JPH09200768A - Moving image processing method and apparatus - Google Patents

Abstract

(57) [Abstract] [PROBLEMS] To enable calculation of an appropriate threshold value without imposing a heavy burden on a user. Kind Code: A1 When a change in video is detected and the amount of data is reduced and output, one or a plurality of types of threshold values used for detecting the change in the video are input, An error rate for each threshold value in a moving image in a stationary state that the user thinks does not change, and a moving image in a moving state that the user thinks are changing, including a threshold value calculation step 1 that is calculated based on the moving image in the moving state. It is possible to determine the optimum threshold value by only adding the error rate for each threshold value in 1) and inputting the moving image in the still state and the moving image in the moving state.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving image processing method and apparatus, and more particularly to a moving image processing method and apparatus, a monitoring apparatus for transmitting and storing images in a monitoring area, a TV conference apparatus, or a general-purpose computer for handling moving images. It is suitable.

[0002]

2. Description of the Related Art Conventionally, in the field of monitoring such as road monitoring and nighttime monitoring, by transmitting a video of a monitoring area photographed by a television camera or the like to a remote place for monitoring or accumulating in a storage device. , Devices that reduce the cost of monitoring are used.

In recent years, such as in TV conferences,
Even without the purpose of monitoring, images are often transmitted and stored between remote places through a communication network (WAN or LAN). Further, as the performance of the general-purpose computer has improved, it is becoming possible to transmit and store the moving image on the general-purpose computer only by displaying the moving image.

Many methods have been proposed in the past for detecting changes in video. As a relatively simple method, a method of detecting a change area by a difference between images has been proposed (Sakuma, Ito, Masuda, "Intruding Object Detection Method Using Difference Between Frames", Technical Report of the Institute of Television Engineers of Japan, vol. .14, No49, pp. 1-6, 1990, etc.).

As a method of detecting a change in video due to the difference between the images, a method of detecting a difference between a background image captured in advance and a current image, a method of detecting a change area based on a difference between adjacent frames, or the like is used. Is a typical method.

As a method of detecting the change area based on the difference between the frames, a device for recording the video of the monitoring area on the VTR detects the change of the video and records the video on the VTR only when the video changes. A device for doing so has been proposed (for example, Japanese Patent Laid-Open No. 2-2486).

Further, the present applicant has already detected the change of the video by comparing the latest image in which the change is detected (the image transmitted, accumulated and displayed) with the current image as shown in FIG. Suggests a way to detect. As a result, it is not necessary to prepare a background image in advance, and it is possible to detect even a gradual change in video.

[0008]

In the above-described image change detecting method, etc., noise in the image data is taken into consideration, and when the amount of change exceeds a certain value (threshold value), it is detected that the image has changed. Is common.

However, the optimum threshold value is often different depending on the image quality (amount of noise) and the content of the scene (brightness of the scene). Further, there are many cases where the threshold value needs to be changed according to the performance of the camera or the A / D converter or the installation location of the camera.

For this reason, conventionally, the task of determining an appropriate threshold value is troublesome, and the user must determine the appropriate threshold value each time by changing the threshold value many times and looking at the image change detection result. A lot of complicated work was required.

In view of the above problems, it is an object of the present invention to calculate and set an appropriate threshold value without imposing a heavy burden on the user.

[0012]

A moving image processing method of the present invention is a moving image processing method for detecting a change in video, reducing the amount of data, and outputting the reduced still image. It is characterized by including a threshold value calculating step of calculating one or a plurality of kinds of threshold values to be used for detecting a video change based on the moving image and the moving image of the moving state.

Another feature of the present invention is that
It is characterized in that one or a plurality of types of threshold values are calculated by creating an error rate at each threshold value from the moving image in the stationary state and the moving image in the moving state.

Another feature of the present invention is that one or a plurality of types of threshold values are calculated by summing up the error rates in the moving image in the stationary state and the moving image in the moving state for each threshold value. It is characterized by that.

Another feature of the present invention is that by weighting the error rates of the moving image in the stationary state and the moving image in the moving state and summing them for each threshold value,
It is characterized in that one or a plurality of types of threshold values, which give priority to prevention of missed detection or erroneous detection, are calculated.

Another feature of the present invention is that the user specifies the priority of the above-mentioned omission of detection or prevention of erroneous detection step by step.

Another feature of the present invention is that one or more kinds of error rates are minimized from the error rate at each threshold value created from the moving image in the stationary state and the moving image in the moving state. It is characterized in that the threshold value of is calculated.

Another feature of the present invention is that the error rate is less than or equal to a certain value from the error rate at each threshold value created from the moving image in the stationary state and the moving image in the moving state. It is characterized in that a range of a plurality of types of thresholds is calculated, and the user specifies one or a plurality of types of thresholds from the range of the calculated thresholds.

Another feature of the present invention is that a threshold value for calculating one or a plurality of types of threshold values used for detecting a video change is calculated based on a moving image in a stationary state and a moving image in a moving state. A calculation step, and an image input step of inputting a predetermined still image as an input image from the video imaged by the imaging means or the video image stored in the recording device,
A change detection step of comparing the input image with a latest change image in which a change has been detected to detect a change in video; a latest change image storage step of storing the latest change image in which the change has been detected; It is characterized by including an image output step of outputting an image in which a change is detected among the images.

Another feature of the present invention is that the threshold value calculating step calculates one or more kinds of threshold values.

A feature of the moving image processing apparatus of the present invention is that the moving image processing apparatus detects a change in a video and outputs the reduced video data amount. It is characterized by comprising a threshold value calculation means for calculating one or a plurality of kinds of threshold values used for detection, based on a moving image in a stationary state and a moving image in a moving state.

Another feature of the moving image processing apparatus of the present invention is that one or a plurality of types of threshold values used for detecting a video change are set to a moving image in a stationary state and a moving image in a moving state. The present invention is characterized by including a threshold value calculation unit that is calculated on the basis of the threshold value, and a priority instruction unit that allows the user to give a stepwise instruction to prioritize prevention of missed detection or false detection.

Another feature of the moving image processing apparatus of the present invention is that it is used for detecting a video change.
One or more kinds of threshold values are calculated based on a moving image in a still state and a moving image in a moving state, and a user specifies a predetermined threshold value from a threshold value range calculated by the threshold value calculating means. It is characterized in that it is provided with a threshold value indicating means for making

Another feature of the moving image processing apparatus of the present invention is that it is used for detecting a video change.
Threshold value calculation means for calculating one or more kinds of threshold values based on a moving image in a still state and a moving image in a moving state, and a still image from a video image captured by the image capturing means or a video stored in a recording device Image input means, a change detection means for detecting a change in video by comparing the input image with a latest change image in which a change has been detected, and a latest change image storage means for storing the latest change image. Of the input images, an image output unit that outputs only an image in which a change is detected is provided.

Another feature of the moving image processing apparatus of the present invention is that the threshold value calculating means calculates one or a plurality of kinds of threshold values.

[0026]

Since the present invention comprises the above technical means, based on the input still image and moving image,
Since one or a plurality of types of thresholds for use in detecting a video change are calculated, the user can calculate and set the optimum threshold by simply inputting a moving image in a still state and a moving image in a moving state. It will be possible.

According to another feature of the present invention, the error rate for each threshold value in a moving image in a stationary state that the user thinks is unchanged, and the threshold value in a moving image in a moving state that the user thinks are changing. One or a plurality of types are calculated by adding the error rate for each of them, that is, the sum of the rate of detecting a change in a still moving image and the rate of not detecting a change in a moving image. Since the threshold value is calculated, it is possible to automatically obtain the threshold value that minimizes the error rate and calculate an appropriate threshold value.

According to another feature of the present invention, when the user designates the false detection prevention priority, the error rate distribution in the stationary state is weighted and added to the error rate distribution in the dynamic state, When the user specifies change detection / leakage prevention priority, one or a plurality of types of threshold values are calculated by weighting the error rate distribution in the moving state rather than the error rate distribution in the stationary state and calculating the threshold value. It is possible to appropriately comply with either the request for prevention of missed detection or the priority for prevention of false detection.

According to another feature of the present invention, the threshold range with a low error rate is presented to the user,
It becomes possible for the user to select the final threshold value, and the user can finely set the designation of the false detection prevention priority and the detection failure prevention priority.

[0030]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of a moving image processing method and apparatus according to the present invention will be described below with reference to the drawings. First,
The first embodiment will be described.

FIG. 1 is a diagram for explaining a typical configuration example in which the present invention is applied to a moving image processing method. In FIG. 1, a threshold calculation step 1 calculates an optimum threshold from a moving image in a stationary state and a moving image in a moving state.

In the image input step 2, a still image (hereinafter referred to as an input image) is obtained from an image obtained by a video camera or an image such as a digital moving image stored in a disc device.

In the change detection step 3, the input image input in the image input step 2 is compared with the latest image in which a change is detected (hereinafter referred to as the latest change image), and the threshold value calculation step is applied to the comparison result. The change in the image is detected using the threshold value calculated in 1.

The change detection determining step 4 is a step of determining whether or not a change in the image is detected in the change detecting step 3. If the change in the image is detected in the change detecting step 3, the latest change image is detected. Transfer processing to storage step 5,
When no change can be detected, the process is returned to the image input process 2.

The latest change image storage step 5 stores the input image as the latest change image when the change is detected in the change detection step 3. In the image output step 6, when a change is detected in the change detection step 3, the input image is output to a communication path (WAN or LAN) and transmitted, or supplied to and accumulated in an image accumulating device, and an image display device. It is supplied to and displayed.

The process end judging step 7 is a step of judging whether or not the user has given an instruction to end the process. If the instruction to end the process has not been given, the process is returned to the image input step 2 described above. If the instruction to end the process is given, the process ends.

FIG. 2 is a functional block diagram showing a configuration example of a moving image processing apparatus for performing the processing described in FIG.
Is an example of hardware that realizes the configuration shown in FIG. In FIG. 3, reference numeral 2001 is a CPU, 2002 is a disk device, 2003 is disk I / O, and 2004 is RO.
M (Read Only Memory), 2005 is a video capture device.

Reference numeral 2006 is a RAM (random access memory), 2007 is a bus connecting the above-mentioned blocks, 2008 is a video camera, and 2009 is a WA.
A communication device that connects to an N (wide area network) or a LAN (local area network).
10 is a VTR device.

In FIG. 2, the image input means 1001 is
Video obtained by video camera 2008 and disk device 2
A still image is acquired from a digital moving image or the like stored in 0002 and stored in the input image storage unit 1004.

For example, in FIG. 3, ROM 2004
Or C operated by a program on RAM 2006
Under the control of the PU 2001, a still image (input image) is acquired by the video capture device 2005 from the video of the video camera 2008 or the VTR device 2010, and the RAM 2
It is stored in 006.

Alternatively, the ROM 2004 or the RAM 20
CP operated by the program stored in 06
Under the control of U2001, a still image (input image) may be read from the digital moving image stored in the disk device 2002 and stored in the RAM 2006.

Although not shown in FIG. 3, the ROM 2
A still image (input image) is captured by a digital still camera under the control of the CPU 2001 operated by a program stored in the RAM 004 or the RAM 2006, and RA
It may be stored in M2006. Also, a plurality of image input means may be provided.

The change detecting means 1002 detects a change in video by comparing the input image stored in the input image storage means 1004 with the latest change image stored in the latest change image storage means 1005.

When the change detecting means 1002 detects a change in the image, the input image accumulated in the input image accumulating means 1004 is accumulated in the latest change image accumulating means 1005. The change detection means 1002 is, for example, in FIG. 3, a CPU 2001 operated by a program on a ROM 2004 or a RAM 2006, a RAM 2006 used as a work memory, or a disk device 20.
02 can be configured. Of course, the CPU, the RAM, and the disk device dedicated to each may be provided and configured, or configured by dedicated hardware.

The image output means 1003 receives the input image stored in the input image storage means 1004 as a communication path (WA) when the change detection means 1002 detects a change in the image.
(N or LAN) for transmission, output to the input image storage device 1004 for storage, or display on an image display device (not shown).

For example, in FIG. 3, ROM 2004
Or C operated by a program on RAM 2006
Under the control of the PU 2001, the data is transmitted to the WAN or LAN through the communication device 2009.

Alternatively, the ROM 2004 or the RAM 20
Under the control of the CPU 2001 operated by the program on the disc 06, as a part of the digital moving image, the disc device 2
It is stored in 002. Here, the disk device 2002 is
It may be available through a network such as a LAN.

Although not shown in FIG. 3, an image output unit for displaying a moving image can also be configured by continuously displaying a still image on the same portion of the display. A plurality of image output means may be provided.

The input image storage means 1004 and the latest change image storage means 1005 are, for example, RAM 2 in FIG.
006 or a disk device 2002. Of course, each unit may be configured by a dedicated storage device.

The threshold calculating means 1006 is a change detecting means 100 when a moving image in a stationary state and a moving image in a moving state are input.
A threshold value is calculated from the detection result of 2. Threshold calculation means 100
6 is a ROM 2004 or RA in FIG.
CPU200 operated by a program on M2006
1 and a RAM 2006 used as a work memory, or a disk device 2002. Also in this case, each may be configured by a dedicated CPU, RAM, disk device, or the like, or may be configured by dedicated hardware.

The steps 1 to 7 of FIG. 1 will be described below with reference to FIG.
It will be described in detail with reference to the configuration diagram of FIG. <Threshold Calculation Step> The threshold calculation step 1 includes the threshold calculation means 10
In 06, it is a step of calculating a threshold value from the detection result of the change detecting means 1002 when the moving image in the stationary state and the moving image in the moving state are input.

FIG. 4 is a detailed block diagram of the threshold value calculation step 1. Further, FIG. 5 shows how the operation of the threshold calculation step 1 is described. In FIG. 4, a still state error rate distribution calculation step 101 inputs a still image (a moving image that the user thinks is unchanged) from the video camera 2008 or the VTR device 2010, and detects change while changing the threshold value. Then, the error rate distribution in the stationary state (at each threshold, the distribution of the ratio of detecting a change in the moving image in the stationary state) is calculated. For example, the distribution of error rates as shown in FIG. 5A is calculated.

In FIG. 4, in the moving state error rate distribution calculating step 102, a moving image in a moving state (a moving image which the user thinks is changing) is input from the video camera 2008 or the VTR device 2010, and the threshold value is changed. While detecting the change, the error rate distribution of the moving state (the distribution of the ratio of not detecting the change with respect to the moving image of the moving state at each threshold) is calculated. For example, the distribution of error rates as shown in FIG. 5B is calculated.

In FIG. 4, the total error rate distribution calculation step 103 is performed by the error rate distribution calculation step 101 in the stationary state and the error rate distribution in the static state and the error rate in the dynamic state calculated in the error rate distribution calculation step 102 in the dynamic state. The distribution is added for each threshold to calculate the total error rate distribution.

For example, the error rate distribution in a static state as shown in FIG. 5A, the error rate distribution in a dynamic state as shown in FIG. 5B, and the total error rate distribution shown in FIG. 5C (thick line) )
Is calculated. Here, an example has been described in which the error rate distribution in the static state and the error rate distribution in the dynamic state are added with the same weight.

However, depending on the purpose of use of the user, when it is desired to give priority to prevention of missed change detection (corresponding to error rate in moving state) over prevention of erroneous detection (corresponding to error rate in stationary state), or vice versa. Can also be assumed.

Therefore, the error rate distribution in the stationary state and the error rate distribution in the moving state may be weighted unevenly and added. Here, the weight to be applied to the error rate distribution is RO
The value written in the M2004 or the RAM 2006 or the disk device 2002 may be used.

Further, the user's erroneous detection prevention priority (corresponding to lowering the error rate in the stationary state) and the change detection omission are prevented by using a user interface which is a combination of a display device and a pointing device (not shown). The weight to be added may be determined based on the designation of priority (corresponding to lowering the error rate in the moving state).

For example, when the user designates the error detection prevention priority, the error rate distribution in the static state is weighted more than the error rate distribution in the dynamic state (the error rate distribution in the static state is doubled, etc.). Etc.). When the user specifies change detection / leakage prevention priority, the error rate distribution in the dynamic state is weighted more than the error rate distribution in the static state and added (double the error rate distribution in the dynamic state, etc. ). Further, the user's erroneous detection prevention priority and detection leakage prevention priority may be designated in stages, and the weight may be changed according to the designated stage.

In FIG. 4, the minimum error rate threshold calculation step 104 calculates the threshold value with the minimum error rate in the total error rate distribution calculated in the total error rate distribution calculation step 103, and sets it as the optimum threshold value. . For example, FIG. 5 (c)
In the case of the total error rate distribution as shown in FIG. 5, the minimum value shown in FIG. 5C is calculated.

Although the threshold value for which the error rate is the minimum is calculated as an appropriate threshold value here, the threshold value range for which the error rate is low (for example, the threshold value for which the error rate is less than a certain value such as twice the minimum error rate, (Range) may be presented to the user, and a display device (not shown) and a pointing device may be combined so that the user selects a final threshold value. In this case, the burden on the user is slightly increased, but the advantage that the user can finely set the designation of the false detection prevention priority and the detection failure prevention priority is obtained.

Here, in order to calculate the error rate distribution in the stationary state in the stationary state error rate distribution calculating step 101 and the moving state error rate distribution in the moving state error rate distribution calculating step 102, the error rate distribution in the ROM 2004 or the RAM 2006 is stored. An example of the procedure executed by the CPU 2001 operated by the program is shown in the flowchart of FIG. Here,
The case where there is one threshold value used in the change detection step 3 described below will be described as an example.

In step S501, a moving image in a stationary state or a moving image in a moving state is input, and with a set threshold value, for example, an image input step 2, a change detecting step 3, and a latest changing image accumulating step 5 will be described. Change detection processing is performed.

Step S502 is an error rate at each threshold.
R is the ratio of detecting the change in the still moving image or the ratio of not detecting the change in the moving image.
The data is stored in the AM 2006 or the disk device 2002.

In step S503, it is determined whether steps S501 to S502 have been performed for all threshold values. Then, when it is determined that the processing is completed with all the threshold values, the processing shown in FIG. 6 is ended. If it is determined that the processing has not been completed for all thresholds, step S5
Go to 04.

In step S504, step S501
After changing (increasing) the threshold value used in the change detection process performed in step S501, the process returns to step S501. Here, an example in which the threshold value is increased by one has been described, but if it is desired to obtain the threshold value at high speed because accuracy may be low, the threshold value may be increased in two or more steps. In this way, the error rate distribution in the static state is calculated in the static state error rate distribution calculating step, and the dynamic state error rate distribution is calculated in the dynamic state error rate distribution calculating step.

As for the initial value of the threshold value, the error rate is measured in the static state error rate distribution calculating step and the dynamic state error rate distribution calculating step before the processing shown in the flowchart of FIG. 6 is started. It shall be set (initialized) to the minimum value in the range of threshold values.

<Image Input Step> In the image input step 2, a video obtained by a video camera, a digital moving image stored in a disk device, a still image from a digital still camera, etc. are obtained and stored in the input image storage means 1004. To do.

<Change Detection Step> In the change detection step 3, the change in the video is detected by comparing the input image stored in the input image storage means 1004 with the latest change image stored in the latest change image storage means 1005. To do. FIG. 7 shows the ROM 2004 or the RAM 20 in the change detection step 3.
An example of the processing procedure executed by the CPU 2001 operated by the program on 06 is shown.

This processing calculates the pixel value difference (absolute value) between corresponding pixels in the input image and the latest change image,
If the sum of the pixel value difference values of the entire image is greater than or equal to a certain value,
This is a process of determining that the input image has changed compared to the latest changed image, that is, a process of determining that the video has changed.

Here, as shown in FIG. 8, the pixel value difference is calculated by processing each pixel of the image in raster order, and the sum of the pixel value differences during the process is called the total change amount. Of course, each pixel may be processed in parallel.

In FIG. 7, in step S101, the total change amount is initialized to zero. In the next step S102, the pixel value difference between the pixels being processed (hereinafter referred to as the pixel of interest) is calculated.

For example, if the input image is a binary image, the pixel value difference is set to 1 if the value of the pixel of interest is different, and is set to 0 if they are the same. If the input image is a grayscale image,
For example, the absolute value of the difference between the values of the pixel of interest is set. Furthermore, if the input image is a color image, for example, the RG of the pixel of interest
The absolute value of each difference of B values is calculated, and the absolute value of each difference is summed.

In step S103, step S102
The pixel value difference calculated in step 3 is added to the total change amount. In step S104, it is determined whether or not the total change amount is larger than a preset value (threshold value). If the total change amount is larger, it is determined that the image has changed, and the flowchart shown in FIG. If it is determined that it is not large, step S1
The process proceeds to 05.

In step S105, it is determined whether or not all pixels have been processed. Then, when all the pixels have been processed, it is determined that there is no change in the image, and the process ends.
In step S106, the pixel to be processed (target pixel)
Move to the next pixel and return to step S102.

<Change Detection Judging Step> Change Detection Judging Step 4
Is a step of determining whether or not a change in video is detected in the change detection step 3, and if a change in video is detected, the process proceeds to the latest change image storage step 5, and the change in video is performed. If is not detected, the process is returned to the image input step 2 described above.

<Latest change image storage step> In the latest change image storage step 5, when the change is detected in the change detection step 3, the input image input in the image input step 2 and stored in the input image storage means 1004 is input. The latest changed image is stored in the latest changed image storage unit 1005.

<Process End Judging Step> Process End Judging Step 7
Is for determining whether or not the user has given an instruction to end the process. If the user has not given an instruction to end the process via the communication device 2009, the process returns to the image input step 2 described above. If the processing end instruction is given, the processing is ended.

<Image Output Step> In the image output step 6, when the change in the image is detected in the change detection step 3, the input image stored in the input image storage means 1004 is transmitted through the communication device 2009 or the like to a communication channel (WAN or WAN). It is output to and transmitted to a LAN), output to and stored in the image storage device 2002, or displayed on a display device.

Next, a second embodiment of the present invention will be described. In the present embodiment, an example of implementation of the threshold value calculation step 1 when the change detection step 3 in the first embodiment is realized by a method using two threshold values will be shown.

FIG. 9 shows the ROM in the change detection step 3.
An example of a processing flow chart executed by the CPU 2001 operated by a program on the 2004 or the RAM 2006 is shown. This process calculates the pixel value difference (absolute value) between corresponding pixels in the input image and the latest change image,
If this pixel value difference is greater than or equal to a certain value, it is assumed that the pixel of interest has changed (hereinafter, the changed pixel is referred to as a changed pixel).

Further, if the number of changed pixels in the input image is greater than or equal to a certain value, it is determined that the input image has changed compared with the latest changed image. That is, it is determined that the image has changed. Note that, as in the first embodiment, the determination of the changed pixels is performed by processing each pixel of the image in raster order as shown in FIG. 8, and the total number of changed pixels in the process is called the number of changed pixels. Of course, each pixel may be processed in parallel.

In FIG. 9, in step S201, the number of changed pixels is initialized to zero. In step S202, the pixel value difference between the pixels being processed (hereinafter referred to as the pixel of interest) is calculated by the same method as in the first embodiment. In step S203, it is determined whether the pixel value difference calculated in step S202 is larger than a preset value (first threshold value). If it is larger than the first threshold value, the process proceeds to step S204. On the contrary, if it is not larger than the first threshold value, the process jumps to step S206.

In step S204, the number of changed pixels is increased by one. Next, in step S205, it is determined whether the number of changed pixels is larger than a preset value (second threshold value). In this case, it is assumed that there is a change in the video, and the processing of the flowchart shown in FIG. 9 ends. If it is determined that the size is not large, the process proceeds to step S206.

A step S206 decides whether or not all the pixels have been processed. When all the pixels have been processed, it is considered that there is no change in the image, and the processing of the flowchart shown in FIG. 7 ends. If all pixels have not been processed, the process proceeds to step S207. Step S20
In step 7, the process is moved to the next pixel (target pixel) and the process returns to step S202.

<Threshold Value Calculating Step> As described above, the threshold value calculating step 1 when the change detecting step 3 is realized by a method using two threshold values will be described below. Even in such a case, the threshold value calculation step 1 can be realized by the method shown in FIG. 4 as in the first embodiment. However, since two thresholds are used, the total error rate distribution shown in FIG.
The distribution is as shown in 0.

Further, the error rate distribution calculation step 1 in the stationary state
The method for calculating the error rate distribution in the stationary state in 01 and the method for calculating the error rate distribution in the moving state in the moving state error rate distribution calculating step 102 are also different. Below, ROM2004
Or C operated by a program on RAM 2006
An example of processing executed by the PU 2001 is shown in the flowchart of FIG.

In FIG. 11, in step S601, a moving image in a stationary state or a moving image in a moving state is input, and change detection processing is performed with a set threshold value. For example, the change detection process including the image input process 2, the change detection process 3, and the latest change image accumulation process 5 described above is performed.

In step S602, the error rate (the rate of detecting a change in a moving image in a stationary state or the rate of not detecting a change in a moving image in a moving state) for each combination of thresholds is set in the RAM 2006 or the disk device 2002.
To store.

In step S603, it is determined whether or not the processes of steps S601 to S602 have all been performed with the first threshold value. If all the processes have been completed with the first threshold value, the process proceeds to step S605, while the first If all the processing is not completed with the threshold value of, the process proceeds to step S604.

In step S604, step S601
The first threshold value used in the change detection process performed in (1) is changed (increased), and then the process proceeds to step S601. Similar to the first embodiment, the step size of the first threshold may be any number.

In step S605, it is determined whether or not the processes of steps S601 to S604 have all been performed with the second threshold value.
The process shown in is ended. If all the processes have not been completed, the process proceeds to step S606.

In step S606, step S601
The first threshold value used in the change detection process performed in step 1 is returned to the original value, the second threshold value is increased, and the process proceeds to step S601. Similarly, the step size of the second threshold may be any number. In this way, the error rate distribution calculation step 101 in the static state calculates the error rate distribution in the static state, and the error rate distribution calculation step 102 in the dynamic state calculates the error rate distribution in the dynamic state.

Note that both the first threshold value and the second threshold value are initialized to the minimum value in the range of the respective measured values in the step before the processing shown in the flowchart of FIG. 11 is executed. Shall be.

Next, a third embodiment of the present invention will be described. In the present embodiment, an example of realizing the threshold value calculating step 1 when the change detecting step 3 in the first embodiment is realized by a method using three threshold values will be shown.

In FIG. 12, in the change detection step 3, RO
An example of a processing procedure executed by the CPU 2001 operated by a program on the M2004 or the RAM 2006 will be shown. In this process, the input image and the latest change image are displayed as shown in FIG.
As shown in FIG. 3, the block is divided into a plurality of blocks, and the change is detected for each block.

In each block, as in the second embodiment,
A pixel value difference (absolute value) between corresponding pixels in the input image and the latest changed image is calculated, and if the pixel value difference is equal to or larger than a certain value, the pixel of interest is changed.

Further, if the number of the changed pixels in each block is a certain value or more, it is determined that the block is changed. Finally, when the number of changed blocks is equal to or larger than a certain value, it is determined that the input image has changed compared with the latest changed image.

That is, it is determined that the image has changed. Calculation of the pixel value difference in each block is performed as shown in FIG.
As shown in (b), each pixel is processed in raster order, and the block is also processed in raster order. Further, the total number of changed pixels in the block in the middle of processing is called the number of changed pixels, and the total number of changed blocks is called the number of changed blocks. Of course, each pixel or block may be processed in parallel.

In FIG. 12, in step S401, the number of changed blocks is initialized to 0. Step S4
In 02, the number of changed pixels is initialized to 0. Step S4
In 03, similarly to the first embodiment, the pixel value difference between the target pixels is calculated.

In step S404, step S403
Pixel value difference calculated in step is a preset value (first threshold)
If it is larger, it is determined that the pixel of interest has changed, and the process proceeds to step S405.
If not, the process jumps to step S407.

In step S405, the number of changed pixels is increased by one. In step S406, it is determined whether or not the number of changed pixels is larger than a preset value (second threshold value), and if larger, it is determined that the target block has changed in step S4.
The process proceeds to 08. If not larger, the process proceeds to step S407.

In step S407, it is determined whether all the pixels in the block have been processed. As a result of this determination, if all the pixels have been processed, the process proceeds to step S412,
If all pixels have not been processed, step S411
Proceed to.

At step S408, 1 is added to the number of changed blocks. In step S409, it is determined whether or not the number of changed blocks is larger than a preset value (third threshold value). If larger, it is determined that the image has changed, and the process of the flowchart shown in FIG. 12 ends. If not larger, the process proceeds to step S410.

In step S410, it is determined whether all blocks have been processed. Then, when all the blocks have been processed, it is considered that there is no change in the video, and the processing of the flowchart shown in FIG. 12 ends. On the other hand, if all blocks have not been processed, step S4
Proceed to 12.

In step S411, the process is moved to the next pixel (target pixel), and then the process returns to step S403. In step S412, the processing is performed (the block of interest)
Move to the next block, and then return to step S402.

<Threshold Calculation Step> As described above,
The threshold value calculating step 1 in the case where the change detecting step 3 is realized by a method using three threshold values will be described below. Even in such a case, the threshold value calculation step 1 can be realized by the method shown in FIG. 4 as in the first embodiment described above.
However, since three threshold values are used as in the second embodiment, the error rate distribution is different from the total error rate distribution shown in FIG.

Here, the diagram of the total error rate distribution is omitted. Further, the method for calculating the error rate distribution in the stationary state in the stationary state error rate distribution calculating step 101 and the method for calculating the moving state error rate distribution in the moving state error rate distribution calculating step 102 are also different. Hereinafter, ROM 2004 or RAM 2006
It is executed by the CPU 2001 operated by the above program.

FIG. 14 shows an example of the processing procedure.
In step S701, a moving image in a still state or a moving image in a moving state is input, and a change is made with a set threshold value, for example, the image input step 2, the change detecting step 3, and the latest change image accumulating step 5 described above. Perform detection processing.

In step S702, the error rate (the rate of detecting a change in a moving image in a stationary state or the rate of not detecting a change in a moving image in a stationary state) for each combination of thresholds is set in the RAM 2006 or the disk device 2002.
To store.

A step S703 decides whether or not the processes of the steps S701 to S702 have all been performed with the first threshold value. If the process is not completed, the process proceeds to a step S704, and if the process is completed, the step S703 is executed. S7
Go to 05.

In step S704, step S701
The first threshold value used in the change detection process performed in (4) is changed (increased) and the process proceeds to step S701. Note that, as in the first embodiment, the step size of the first threshold may be any number.

A step S705 decides whether or not the processes of the steps S701 to S704 have all been carried out with the second threshold value, and if the processes are completed, a step S707.
If not, the process proceeds to step S706 if the process is not finished.
Proceed to.

In step S706, step S701
Return the first threshold used in the change detection processing in
The second threshold value is increased and the process proceeds to step S701. Similarly, the step size of the second threshold may be any number.

A step S707 decides whether or not the processes of the steps S701 to S706 have all been carried out with the third threshold value. If the process is finished, the process shown in FIG. 14 is finished, and if the process is not finished. Is step S70
Proceed to 8.

In step S708, step S701
The first threshold value and the second threshold value used in the change detection process performed in step 3 are restored, the third threshold value is increased, and the process proceeds to step S701. Similarly, the step size of the third threshold may be any number. In this way, the error rate distribution calculation step 101 in the static state calculates the error rate distribution in the static state, and the error rate distribution calculation step 102 in the dynamic state calculates the error rate distribution in the dynamic state.

The first threshold value, the second threshold value,
It is assumed that all of the third threshold values are initialized to the minimum value in the range of the respective measured values in the step before the processing shown in the flowchart of FIG. 14 is executed.

[0118]

As described above, according to the present invention, one or a plurality of types for use in detecting a video change are input based on an input still-state moving image and a moving-state moving image. Since the threshold value is calculated, the user can calculate and set the optimum threshold value only by inputting the moving image in the still state and the moving image in the moving state.

Further, according to another feature of the present invention, the error rate for each threshold value in a moving image in a stationary state that the user thinks is unchanged, and the threshold value in a moving image in a moving state that the user thinks are changing. One or a plurality of types are calculated by adding the error rate for each of them, that is, the sum of the rate of detecting a change in a still moving image and the rate of not detecting a change in a moving image. Since the threshold value is calculated, it is possible to automatically obtain the threshold value that minimizes the error rate and calculate an appropriate threshold value.

According to another feature of the present invention, when the user designates the false detection prevention priority, the error rate distribution in the stationary state is weighted and added to the error rate distribution in the dynamic state, When the user specifies the change detection / leakage prevention priority, the error rate distribution in the dynamic state is weighted and added to the error rate distribution in the stationary state to calculate one or more kinds of threshold values. Therefore, it is possible to meet the request of either the detection omission prevention priority or the false detection prevention priority.

According to another feature of the present invention, the threshold range with a low error rate is presented to the user,
The final threshold value can be selected by the user, and the user can finely set the designation of the false detection prevention priority and the detection failure prevention priority.

[Brief description of drawings]

FIG. 1 is a flowchart showing a typical processing procedure of the present invention.

FIG. 2 is a block diagram showing a configuration example of a moving image processing apparatus to which the present invention has been applied.

FIG. 3 is a configuration diagram showing a hardware implementation example of a moving image processing apparatus to which the present invention is applied.

FIG. 4 is a flowchart showing an example of detailed processing of a threshold value calculation step.

FIG. 5 is a characteristic diagram showing an example of an error rate distribution.

FIG. 6 is a flowchart showing an example of processing of an error rate distribution calculating step in a stationary state or an error rate distribution calculating step in a moving state.

FIG. 7 is a flowchart showing a process of a change detection process.

FIG. 8 is a diagram illustrating a pixel processing order.

FIG. 9 is a flowchart illustrating a process of a change detection process.

FIG. 10 is a diagram showing an example of an error rate distribution.

FIG. 11 is a flowchart showing an example of processing of an error rate distribution calculating step in a stationary state or an error rate distribution calculating step in a moving state.

FIG. 12 is a flowchart illustrating an example of processing of a change detection process.

FIG. 13 is a diagram illustrating an example of block division, a pixel processing order, and a block processing order.

FIG. 14 is a flowchart illustrating an example of processing of a change detection process.

FIG. 15 is a diagram illustrating an outline of a change detection method.

[Explanation of symbols]

 1 threshold calculation step 2 image input step 3 change detection step 4 latest change image accumulation step 5 image output step 101 static state error rate distribution calculation step 102 dynamic state error rate distribution calculation step 103 total error rate distribution calculation step 104 minimum error Rate threshold calculation step 1001 Image input means 1002 Change detection means 1003 Image output means 1004 Input image storage means 1005 Latest changed image storage means 1006 Threshold calculation means

Claims (14)

[Claims] 1. A moving image processing method for detecting a change in video, reducing the amount of data and outputting the data, wherein one or a plurality of threshold values used for detecting the change in the video are set. A moving image processing method, comprising: a moving image in a stationary state; and a threshold value calculating step for calculating based on the moving image in a moving state. 2. The one or more kinds of threshold values are calculated by creating an error rate at each threshold value from the stationary moving image and the moving image in the moving state. The moving image processing method described. 3. The one or a plurality of types of threshold values are calculated by summing up the error rates of the moving image in the stationary state and the moving image in the moving state for each threshold value. Or the moving image processing method according to any one of 2). 4. One or a plurality of methods giving priority to prevention of missed detection or erroneous detection by weighting the error rates in the moving image in the still state and the moving image in the moving state and summing the error rates for each threshold value. The moving image processing method according to any one of claims 1 to 3, wherein the type threshold value is calculated. 5. The moving image processing method according to claim 4, wherein the user specifies the priority of the detection omission prevention or the false detection prevention step by step. 6. One or a plurality of types of threshold values having a minimum error rate are calculated from the error rate at each threshold value created from the moving image in the stationary state and the moving image in the moving state. The moving image processing method according to any one of claims 1 to 5. 7. A range of one or a plurality of types of threshold values having an error rate equal to or less than a certain value is calculated from the error rate at each threshold value created from the still state moving image and the moving state moving image, 7. The moving image processing method according to claim 1, wherein the user specifies one or a plurality of types of threshold values from the calculated threshold value range. 8. A threshold value calculating step of calculating one or more kinds of threshold values for use in detecting a video change based on a moving image in a stationary state and a moving image in a moving state, and an image supplied from a video source. From the above, a predetermined still image is acquired, and an image input step of inputting the acquired image as an input image, and the input image and the latest change image in which a change is detected are compared to detect a change in video. A change detection step, a latest change image storage step of storing the latest change image in which the change is detected, and an image output step of outputting an image in which a change is detected among the input images. Video processing method. 9. The moving image processing method according to claim 8, wherein the threshold value calculating step calculates one or more kinds of threshold values. 10. A moving image processing apparatus for detecting a change in video, reducing the amount of data and outputting the data, wherein one or a plurality of types of threshold values for use in detecting the change in the video are provided. A moving image processing apparatus, comprising: a moving image in a stationary state and a threshold value calculating means for calculating the moving image in the moving state. 11. A moving image processing apparatus for detecting a change in video, reducing the amount of data and outputting the data, wherein one or a plurality of types of threshold values used for detecting the change in the video are static. A moving image of the state and a threshold value calculating unit that calculates based on the moving image of the moving state, and a priority instructing unit that allows the user to stepwise give priority to the detection failure prevention or the false detection prevention. And a moving image processing device. 12. A moving image processing apparatus for detecting a change in video, reducing the amount of data and outputting the data, wherein one or a plurality of types of threshold values used for detecting the change in the video are static. A moving image of a state and a threshold calculating unit that calculates based on the moving image of the moving state; and a threshold instructing unit that allows the user to instruct a predetermined threshold from the range of the threshold calculated by the threshold calculating unit. A moving image processing device characterized by the above. 13. A moving image processing apparatus for detecting a change in video, reducing the amount of data and outputting the data, wherein one or a plurality of types of threshold values used for detecting the change in the video are static. State moving image and threshold value calculating means for calculating based on the moving state moving image, image input means for inputting a still image from the image supplied from the image supply source, the input image, and the latest change detecting the change A change detecting unit that detects a change in video by comparing the image, a latest change image storage unit that stores the latest change image, and an image in which a change is detected among the input images is output. And an image output unit for performing the moving image processing. 14. The threshold value calculating means calculates one or a plurality of kinds of threshold values.
The moving image processing apparatus according to any one of 3 above.