JPH03260877A - Binarization processing method - Google Patents

Abstract

PURPOSE:To enable a processing at real time by calculating surrounding brightness from a stop value signal, estimating the contrast of a picture signal and exerting the binarizing processing onto the picture signal, to which spatial differentiation processing is executed, based on a threshold value decided by this contrast. CONSTITUTION:The image of a vehicle 5 to pass on a road surface 6 is picked up with the fixed brightness from an automatic iris lens part 2 of a television camera 1, and a picture signal (a) and a stop value signal (b) are outputted. While inputting these picture signal (a) and stop value signal (b), a picture processing part 3 executes the spatial differentiation processing. Together with this processing, the surrounding brightness is calculated from a voltage value to be generated by the stop value signal (b) and a threshold value is decided by estimating the contrast of the picture signal from this calculated surrounding rightness. Based on this threshold value, the binarizing processing is exerted onto the picture signal to which the spatial differentiation processing is executed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a method for binarizing images of vehicles, etc., which are photographed by a television camera installed outdoors.

[Prior Art] Generally, traffic flow measuring devices using television cameras are used in outdoor environments, and therefore, the captured images are greatly affected by changes in weather, ambient brightness, and the like. Conventionally, an auto iris was used in the lens system to maintain the average brightness level of the image at a constant value in order to capture a stable image that would be viewed by a human on a monitor TV. . Therefore, the optimal threshold value for binarization processing for various image conditions can be determined by determining the density distribution of the image and using statistical means based on this tendency. For example, mode method 1 discriminant analysis binary method and P
- Various methods such as the tile method have been proposed.

[Issue Il to be Solved by the Invention] The conventional traffic flow measurement device described above maintains the average brightness level of images at a constant value, so in the case of images taken of the same subject, On cloudy days, even if the average density level is the same, there is a difference in contrast, and this contrast difference has a large effect when extracting the outline of an object by performing differential processing on the captured image. As a result, the outline of the object can be clearly extracted in images with strong contrast such as on a clear day, but in the case of images with weak contrast, the outline cannot be extracted clearly, and in the subsequent binarization process, There was a problem in that the outline could not be extracted clearly unless an optimal threshold value was determined according to each situation.

Furthermore, the threshold value determination process for this binarization process takes a long time, and therefore lacks real-time performance.

[Means for Solving the Problems] In order to solve such problems, the binarization processing method according to the present invention inputs a photographed image signal and performs spatial differential processing, and also uses a method installed in a television camera. The surrounding brightness is calculated from the voltage value generated by the aperture value signal from the auto iris lens section, and the contrast of the image signal is estimated from the calculated surrounding brightness to determine the threshold value.
Based on this threshold value, the image signal subjected to the spatial differential processing is binarized.

[Operation] The surrounding brightness is calculated from the aperture value signal, the contrast of the image signal is estimated, and as a result, the image signal subjected to the spatial differentiation process is binarized based on the determined threshold value.

[Example] Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram of a system to which the binarization processing method of the present invention is applied. In the figure, 1 is a television camera, 2 is an auto iris lens unit, 3 is an image processing device, 4 is a monitor television, 5 is a vehicle, 6 is a road surface, a is an image signal, and b is an aperture value F at the time of image shooting. The aperture value signal converted into a voltage value, C is an output signal.

The television camera 1 is installed so as to photograph the road surface 6 from above, and the vehicle 5 passing on the road surface 6 is photographed with a constant brightness by the auto iris lens section 2 of the television camera 1, and as a result, An image signal a and an aperture value signal S are output. In addition, the image processing section 3 also uses this image signal a.
and the aperture value signal, perform these arithmetic processes, send the results to the monitor TV 4 and display them, and also calculate the traffic flow parameters (number of passing vehicles, vehicle speed, and vehicle space or time). occupancy rate, etc.) and outputs the result as an output signal C.

SUMMARY OF THE INVENTION In order to solve the conventional drawbacks, the present invention aims to determine an optimal threshold value in real time using a simple horizontal aperture. This objective is achieved by estimating the contrast of the image taken. This contrast estimation method will be described in detail below.

In general, the contrast is 0, which depends on the lighting environment of the object.
That is, for example, when the object is illuminated with highly directional light, such as in the case of illumination from a point light source, an image with strong contrast is obtained because the object is strongly shaded. On the other hand, when the object is illuminated with weakly directional light, such as when it is illuminated from all around, an image with weak contrast is obtained because the shadow of the object is weak.

If we apply this situation to an outdoor environment, images with strong contrast will be obtained when the sky is clear and bright (when there are few clouds and the sun can be seen), and images with strong contrast will be obtained when the weather is cloudy and dark.
When the sun is not visible due to cloud cover, images with weak contrast are obtained.

Generally, it is known that when the sky is clear, the sky illuminance (surrounding brightness) is high, and when the sky is cloudy, the sky illuminance is low. This is an attempt to estimate it from the surrounding brightness.

Next, FIG. 2 is a flowchart explaining the operation of the system to which the binarization processing method of the present invention is applied, and FIG. 3 shows the aperture value F and the corresponding voltage value V, that is, the aperture value signal S. A graph showing the relationship, FIG. 4, is an explanatory diagram illustrating the method of spatial differentiation. The operation of this system will be explained based on FIGS. 2 to 4.

First, in step 100 of the flowchart in FIG. 2, a photograph of the vehicle 5 traveling on the road surface 6 is taken by the television camera 1, and then in step 101 it is determined whether the image signal a has arrived. Then, when this is r Y J, this image signal a is input in step 102, and in step 103, this analog signal image signal a is AD converted into a digital signal, and in step 104, spatial differentiation processing is performed to obtain this image signal. The outline of the digitalized signal, that is, the image of the vehicle 5 is extracted.The method of this spatial differentiation processing is not particularly limited, but after this spatial differentiation processing, the absolute value of each image unit is taken, and then the method described below is performed. Binarization processing is performed in step 111 using the optimal threshold value.

In general, spatial differential processing is performed for high spatial frequencies (
This is a process that emphasizes areas with large brightness changes.As mentioned above, this process allows the outline of an object to be extracted, and in the case of digital image processing, a new pixel value can be determined from the difference between neighboring pixels. I can do it.

In the case of this embodiment, this spatial differentiation processing is performed as follows. That is, as shown in the explanatory diagram of the horizontal differentiation process in FIG. 4(a), the data of the n-th pixel is
If the data of n-2>pixel is C, the value of pixel data B is given by B=IC-A (1), and if this operation is performed for all pixels, the horizontal differentiation process is I can do it.

Next, in the same way, the vertical differential processing is performed such that the value of the pixel data B'' is B' = lC', as shown in FIG. 4(b).
-A' l (2), and if this calculation is performed for all pixels, differential processing in the vertical direction can be performed. In this embodiment, horizontal differentiation is performed alternately for odd fields, and vertical differentiation is performed for even fields.

In addition, in the case of this embodiment, performing the above-mentioned logical operations after storing each pixel data in the memory requires memory transfer time, which affects the processing speed. Processing is done in real time so as not to affect the That is, the current pixel data and the past pixel data C temporarily stored in the memory are read out, a logical operation is performed, and the pixel data Bt!
- It is structured so that it can be obtained.

Next, the aperture color signal will be explained. Simultaneously with the arrival of the above image signal a, as shown in the graph of FIG.
is converted to a voltage value V and sent out as an aperture color signal,
This transmitted aperture color signal S is determined to be "Y" in step 105, and this aperture color signal S is inputted in step 106. Then, CPtJ (not shown) reads the voltage value of this aperture color signal, and in step 108
In step 109, the surrounding brightness is calculated from the read voltage value, and in step 109, the contrast of all input image signals a is estimated.

Then, a threshold value is calculated from this estimated contrast, and in step 111, the image subjected to the spatial differentiation processing is binarized based on the calculated threshold value. If it is determined in step 105 that the aperture color signal cannot be received and the result is rN, the process returns to step 100 and vehicle photography is performed.

Here, this binarization processing is performed using a lookup table method. That is, the lookup table is
This is for performing Ws tone conversion by hardware, and by giving threshold data to the address port of the memory, the image data stored inside the memory is outputted as new data. For example, the threshold value is 10.
When binarizing with
Write "1" in addresses 0 to 255, and if the given threshold data is less than 10, write "1".
0" is output, and when it is greater than 10, "1" is output.

Next, regarding the process of image processing, see (a) to (c) in Figure 5.
This will be explained in detail with reference to the drawings. In the figure (a), the input original image is a grayscale image. Figure (b) shows an image for one line in the dotted line portion in the figure. In the figure (b), e is a level signal when the contrast is strong, and f is a level signal (bold line part) when the contrast is weak. Then, the image signal as shown in FIG. 3(b) is subjected to differential processing and the absolute value is taken, resulting in an image signal as shown in FIG. 12(c). (c
) In the figure, g is a level signal when contrast is strong, h is a level signal when contrast is weak (thick line part), j is a threshold when contrast is strong, and k is a threshold when contrast is weak.

In this way, there will be a difference in the shading level of the image after the spatial differential processing when the contrast is strong and when the contrast is weak, so it is necessary to vary the threshold value when performing optimal binarization processing.

Hereinafter, a method of estimating the contrast state for determining the binarization threshold from the surrounding brightness (sky illuminance) will be described.

There is a method of measuring the sky illuminance using an illuminance meter as a means of detecting the surrounding brightness, but in this example,
As described above, the aperture value signal of the auto iris lens unit 2 used in the television camera 1 is used. Since this aperture value signal corresponds to the average brightness of the subject, it has a proportional relationship with the surrounding brightness, that is, the sky illuminance. In the case of this embodiment, the subjects are the vehicle 5, the road surface 6, and other scenery.

In addition, since various vehicles 5 pass on the road surface 6, this has a considerable influence on the average brightness of the subject, and this slows down the response speed of the aperture value signal control system.
Even if the light passes through the lens, it does not affect the average brightness of the subject.

As mentioned above, the relationship between ambient brightness (sky illuminance) and weather conditions is that when the sky is clear, it is bright (sky illuminance is high).
, when the sky is cloudy, it is dark (sky illuminance is low). When the sky is clear, the directionality of light is strong, so the shadows of objects are strong, and when the sky is cloudy, the contrast is weak. This relationship is summarized in the table below.

Therefore, by estimating the contrast of an image from the aperture value signal, it is possible to set an optimal threshold for binarization processing.

As explained above, the binarization processing method of the present invention is
The surrounding brightness is calculated from the aperture value signal of the auto iris lens unit 2, the contrast of the image is estimated from this surrounding brightness, and the optimal threshold value according to the outdoor weather condition is determined based on this estimated value. In addition, since this threshold value is configured to be variable, stable traffic flow measurement can be performed without being affected by outdoor weather conditions.

[Effects of the Invention] As explained above, the binarization processing method according to the present invention estimates the contrast of the image signal by calculating the brightness of the IR area from the aperture value signal, and uses the threshold value determined based on this contrast. Since the system is configured to binarize the image signal that has been spatially differentiated based on the above, the optimal threshold value can be determined according to each situation in the binarization process, and this binarization process can be performed in real time. It has the effect of being able to stably measure traffic flow regardless of outdoor weather conditions.

[Brief explanation of drawings]

Fig. 1 shows the constituent countries showing an embodiment of the system to which the binarization processing method of the present invention is applied, Fig. 2 is a flowchart explaining this operation, and Fig. 3 shows the aperture value F and the corresponding voltage value. A graph showing the relationship with V, FIG. 4 is an explanatory diagram explaining spatial differentiation, and FIG. 5 is an explanatory diagram explaining the process of image processing. l...Television camera, 2...Auto iris lens section, 3...Image processing device, 4...Monitor TV, 5...Vehicle, 6...Road surface, a・・・・・・
Image signal, b...aperture value signal, C...output signal.

Claims (1)

[Scope of Claims] A system that captures images of vehicles passing on a road surface using a television camera to generate an image signal, and processes the generated image signal to measure traffic flow, the system comprising: At the same time, the ambient brightness is calculated from the voltage value generated by the aperture value signal from the auto iris lens provided in the television camera, and the image signal is calculated from the calculated ambient brightness. A binarization processing method, characterized in that a threshold value is determined by estimating the contrast of the image signal, and the image signal subjected to the spatial differential processing is processed based on the threshold value.