JPS621085A - Picture segmenting device - Google Patents

Abstract

PURPOSE:To attain the segmentation of an object picture with simple hard constitution by producing a window pattern along an area within a frame and fetching only the picture input corresponding to a window area. CONSTITUTION:A TV camera 1 photographs an object 2 to be recognized and a synchronizing separator 3 delivers video signals to a binary-coding circuit 5. A mode changeover switch SW is set at the learning mode side for image pick-up of only a frame body 4 of a pallet, etc. Thus a window pattern WP is stored in a window memory 6. Then the SW is set at the measurement mode side for image pick-up of both the body 4 and the object 2. Thus an input pattern Pi including the picture patterns P1 and P2 of the object 2 and the body 4 is obtained. The pattern WP is read out of the memory 6 and applied to a gate circuit 12 as the input to be given to the other side. Then only the constitution data on the pattern P1 included in a window area (w) is sent to a black picture element detecting circuit 13.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field of the Invention The present invention relates to a visual recognition technology that recognizes an object by converting it into an image, for example. The present invention relates to an image cutting device for selecting and capturing only images.

<Summary of the Invention> In the present invention, when an object to be recognized is located in a frame area, first, a model of the frame is imaged to generate a window pattern along the frame area, and then a window pattern is generated along the frame area. The system is configured to capture and input the image of the object along with the frame, and at the same time read out the window pattern and capture only the image input corresponding to the window area.This makes it possible to cut out object images at high speed with a simple hardware configuration.

<Background of the Invention> For example, when visual recognition technology is introduced into a system that stores and transports objects to be recognized in a frame such as a pallet, after cutting out only the image of the object to be recognized, It is necessary to measure the center of gravity, etc., and move on to matching processing for object recognition. In this case, conventionally, the entire frame including the object to be recognized is imaged with a television camera, and as shown in FIG. 5, a window area is created between the image pattern P+ of the object to be recognized and the image pattern P2 of the frame. W is set, and this window is used to cut out only the image pattern P of the object to be recognized. However, in this method, a complicated hardware configuration is required to set the window, and not only does it take time to set the window, but the gap between the image pattern P2 of the frame and the image pattern P1 of the object to be recognized is large. If the window is narrow, there are problems such as difficulty in setting a window with a complicated shape.

<Object of the invention> This invention is intended to solve the above problems,
It is an object of the present invention to provide an image cutting device which has a simple hardware configuration and can cut out an object image by setting a window of an arbitrary shape at high speed.

<Configuration and Effects of the Invention> In order to achieve the above object, the present invention provides an apparatus for image-forming an object to be recognized located in a frame area integrally with the frame, and then cutting out only an image of the object to be recognized. a window pattern generation means for capturing an image of the model of the frame and generating a window pattern along the area within the frame;
a pattern storage means for storing the generated window pattern; and a pattern storage means for capturing and inputting an image of the object to be recognized located in the area within the frame together with the frame, and at the same time reading out the window pattern from the storage means and storing the image in the area of the window pattern. It was decided to include an image capture means for capturing only the corresponding image input.

According to this invention, it is possible to set a window at high speed with a simple hardware configuration, and even if the gap between the image pattern of the frame and the image pattern of the object to be recognized is narrow, a window with a complex shape can be set. can be set, and it is possible to accurately cut out an object image, achieving the remarkable effects of achieving the purpose of the invention.

<Description of Embodiments> FIG. 1 shows an example of the circuit configuration of an image cutting device according to the present invention. In the illustrated example, the television camera 1 images a stationary or moving object 2 from above, for example, and sends an image output (shown in FIG. 2 (11)) related to interlaced scanning to the synchronization separation circuit 3. The recognition object 2 is housed in a frame area of a frame body 4 such as a pallet, and in this device example, the entire image-containing cover turn P including the frame body,
After determining the image pattern (shown in the figure), a window described later is set thereto, thereby cutting out only the image pattern P1 of the object to be recognized.

The synchronization separation circuit 3 receives a horizontal synchronization signal HD from the image output.
, vertical synchronization signal VD, odd field signal OD (shown in FIG. 2 (2)), clock signal CK (shown in FIG. 2 (4)), etc., and outputs the video signal VDi to the binarization circuit 4. do. As shown in FIG. 2 (3), the binarization circuit 4 sets a constant threshold level T to the video signal VDi.
H is set, and the odd fields of the video signal VDi are converted into black and white binarized to form a binary pattern and output.

A window memory 6 is connected to the binarization circuit 5 via a mode changeover switch SW, and when the mode changeover switch SW is set to the learning mode side a, a frame 4 such as a pallet etc.
When capturing only the third image, for example, the third
When the window pattern WP shown in the figure is stored and the mode selector switch SW is set to the measurement mode side to image the object 2 to be recognized together with the frame 4, the input cover turn P shown in FIG. 5 is generated. Ru. In the case of this embodiment, each pattern is assembled into an image range of 256 bits in length and width, and in FIGS. 3 and 5, the shaded areas represent black pixels, and the areas other than the shaded areas represent white pixels, respectively. .

In FIG. 1, the horizontal counter 7 and the vertical counter 8 are
When reading and writing the window pattern WP, pixel positions within the window memory 6 are addressed. Gate circuits 9 and 10 are controlled to open and close by odd field signal OD and clock signal CK, and supply write control signal W and read control signal R to window memory 6. Furthermore, gate circuit 11 is controlled to open and close by odd field signal OD, and supplies clock signal CK to horizontal counter 7 and vertical counter 8, respectively.

A black pixel detection circuit 13 for detecting black pixels in the pattern is connected to the measurement mode b side of the mode changeover switch SW via a gate circuit 12. is connected to the readout output of the window memory 6.

This gate circuit 12 is for setting the window pattern WP stored in the window memory 6 for the input cover pattern P, and therefore the black pixel detection circuit 13 is for setting the window pattern WP stored in the window memory 6 for the input cover pattern P.
The black pixels of the input cover turn P1 included in the recognition object 2, that is, the black pixels constituting the input cover turn P (shown in FIG. 4) on the object to be recognized 2, are detected and the detection signal is sent to the pixel counter 14. The pixel counter 14 counts the output (black pixel number) of the black pixel detection circuit 13, and the count data of this pixel counter 14 is sent via the 1/O (Input/0putput) port 15 every horizontal blanking period. CPU (Central Process)
ssing Unit) 16, and CPU1
6 calculates and outputs the cumulative value (corresponding to the area of the pattern) of the captured count data.

In addition, in the figure, F ROM (Programmable
ReadOnly Memory) 17 stores a series of programs, and RAM (Rando Memory) 17 stores a series of programs.
m Access Memory) 18 not only stores various data but also has a work area for processing execution. Further, the gate circuit 19 is a circuit that generates an interrupt signal INT to the CPU 16, and the OR circuit 20 is a circuit that resets the pixel power counter 4. Furthermore, display section 2
1 displays the cumulative calculation result of the number of black pixels, and a setting switch 22 sets various threshold values and the like.

After setting the mode changeover switch SW to the learning mode a side, when the frame 4 with the recognized object 2 removed is imaged by the television camera l, the first odd field of the video signal VD is binarized. is executed to generate an image pattern of only the frame 4, and this pattern is stored in the window memory 6 as a window pattern WP shown in FIG.

Next, after switching the mode changeover switch SW to the measurement mode b side, when the object 2 to be recognized positioned within the frame on the frame body 4 is imaged by the television camera l, two images are captured for each first odd field of the video signal VD. The value processing is performed to generate an input cover turn Pi including an image pattern P of the object to be recognized 2 and an image pattern P2 of the frame 4 as shown in FIG. This input cover pattern P1 is input to the gate circuit 12, but at the same time, the window pattern WP is read out from the window memory 6 and given as the other input to the gate circuit 12, thereby causing the window area W Only the input pattern P1 included in the image pattern P1, that is, the configuration data of the image pattern P1 related to the object to be recognized 2 passes through the gate circuit 12 and is sent to the black pixel detection circuit 13.

In this way, the black pixel detection circuit 13 detects black pixels included in the window area W, and the next pixel counter 14 receives the detection signal and counts the number of black pixels. Then, at each time timing of the horizontal synchronization signal HD in the odd field, an interrupt signal INT is generated to the CPUI 6,
During this horizontal synchronization period, the CPU 16 takes in the count data of the pixel counter 14 and executes its cumulative calculation. By repeating this calculation and processing for each horizontal scanning line, the image pattern P on the object to be recognized 2 is
Total number of black pixels constituting l (corresponds to the area of pattern P)
is calculated, and the calculation result is output to the display unit 21 or the like.

In the above embodiment, the window pattern WP with white pixels inside the window area W is stored in the window memory 6, but the present invention is not limited to this, and a window pattern with black pixels inside the window area W can be stored in the window memory 6. It may be stored, and in this case, an AND circuit is used as the gate circuit 12.

[Brief explanation of the drawing]

FIG. 1 is a circuit block diagram of an image cutting device according to the present invention, FIG. 2 is a time chart showing signal waveforms of the circuit in FIG. 1, FIG. 3 is a diagram showing window patterns, and FIG. 4 is a diagram showing window setting states. The figure shown in FIG. 5 is a diagram showing an input cover turn. 1...TV camera 5...Binarization circuit 6
...Window memory 12...Gate circuit 1
6...CPU

Claims (2)

[Claims] (1) A device for image-forming an object to be recognized located in a region within a frame, integrally with the frame, and then cutting out only an image of the object to be recognized, which images a model of the frame and images the region within the frame. a window pattern generation means for generating a window pattern along the frame; a pattern storage means for storing the generated window pattern; An image cutting device comprising: image capturing means for reading a window pattern from the storage means and capturing only an image input corresponding to an area of the window pattern. (2) An image cutting device in which the window pattern generation means includes a television camera and a binarization circuit that binarizes an input image.