JPS63318676A - Image recognizing device - Google Patents

Abstract

PURPOSE:To shorten a recognizing processing time and to save a memory capacity by setting a window and a shifting quantity at the time teaching and shifting and collating displaying image data and registering image data at every setting window at the time of recognition. CONSTITUTION:First, at the time of teaching, the video signal of an objective work by a TV camera 1 is subjected to binarization by an image input circuit 2 and the data are displayed by a monitoring TV 4. At such a time, a light pen is operated, the image data of a work are registered and a window 8a and a shifting quantity to show the range of the data reading at the time of recognition, are set. Next, at the time of the recognition mode, the image corresponding to the work is displayed at a TV 4, a first recognition instruction is generated, and then, an image processing part executes the recognizing processing with the position and width of the window 8a and the result is displayed at the TV 4. By a continuous second recognition instruction, the image processing circuit shifts the designating quantity of the window 8a, executes the recognizing processing with the position and the width of a window 8b and the result is displayed at the TV 4. Hereinafter in the same way, the processing is repeated.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an image recognition device suitable for parts supply work, transportation work, positioning work, etc. by automatic machines, robots, etc. in flexible automation (FA). be.

[Conventional technology]

Fig. 5 is a block diagram showing the configuration of this type of conventional image recognition device, in which (1) is a television camera that photographs the recognition target, and (2) is a video signal binarization of the television camera (1). (3) is an image processing circuit that includes a storage device for storing image data and a microcombination λ-2, and performs processing for registering and recognizing image data; (4) is a monitor TV that inputs and displays data from the image processing circuit (3), (5) is a light ben that registers the image data displayed on the monitor TV, (6) is a recognition target (hereinafter referred to as target work), (7) is a container into which the target work (6) is placed, and (8) is a window for specifying the image reading range.

Next, the operation will be explained. Various data necessary for recognizing the target workpiece (6) in the teaching mode are input and registered in the storage device of the image processing circuit (3). Now, when a recognition and identification command is given to the image processing circuit (3), the television camera (1) photographs the target workpiece (6) contained in the container (7), and the video signal is sent to the image input circuit (2). ). The image input circuit (2) binarizes this video signal, or? After being subjected to IA light processing and input to the image processing circuit (3), this signal is applied to the monitor television (4) and displayed inside the window (8) designated by the light ben (5) during the teaching mode.

Next, the image processing circuit (3) compares the image data displayed on the monitor television (4) with various data registered during the teaching operation, and determines the position of the target workpiece (6).
Recognize and identify the tilt, etc. and display the results on a monitor TV (
4).

[Problem that the invention seeks to solve]

In the conventional image recognition device described above, the window (
8) is constant in size, and moreover, the storage device of the image processing circuit (3) often only holds data for one page. Therefore, when images of multiple target works (6) are displayed inside the window (8), the amount of data processed by the image processing circuit (3) becomes enormous, and recognition processing takes a lot of time. There was a problem in that it lacked practicality because it was expensive. Therefore, a method has been proposed in which the size of the window (8) is changed at any time and data corresponding to multiple windows is registered. There were problems in that it was time-consuming and required extremely large memory capacity.

The present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide an image recognition device that can shorten recognition processing time and save memory capacity.

Means for Solving Problem C] The image recognition device according to the present invention performs processing such as binarization on a video signal obtained by photographing a target work with a television camera and displays it on a monitor television, and performs teaching. In the mode, the registration means stores the image data displayed on the monitor TV in the storage device, and also sets a window indicating the range of data reading at the time of recognition and its shift amount,
In the recognition mode, the recognition processing means collates the image data displayed on the television monitor with the image data registered in the storage device for each window that is shifted sequentially, and displays the collation results on the television monitor. be.

[For production]

In this invention, a window indicating the range of data reading for recognition and its shift amount are set for the image displayed on the monitor TV during the teaching mode, and the window indicating the range of data reading for recognition is set for the image displayed on the monitor TV during the recognition mode. Since the data and the image data registered in the storage device are compared while being shifted sequentially for each set window, it is possible to shorten recognition processing time and save memory capacity.

(Embodiment) FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, in which a video signal from a television camera (1) is transmitted to an image input circuit (
2), the image is binarized or gradated and input to a monitor television (4), where the image is displayed. The registration means (10) stores (hereinafter referred to as registration) the image data displayed on the monitor television (4) in the teaching mode in the storage device (9), but at this time, the registration means (lO) Set the window that indicates the data reading range and its shift amount. In addition, image recognition means (11
) compares the image data in the recognition mode displayed on the monitor television (4) with the image data registered in the storage device (9) for each window set by the registration means, and displays the results on the monitor television ( 4) will be displayed.

FIG. 2 is a diagram showing a specific configuration of this embodiment, and the same reference numerals as in FIG. 5 showing the conventional device indicate the same elements. The internal structure of the image processing circuit (3a) consisting of a storage device and a microcomputer is different from that of conventional devices. The detailed operation of this embodiment will be explained below.

First, the video signal of the target workpiece (6) photographed by the television camera (1) during the teaching mode is binarized by the image input circuit (2), and when the obtained data is input to the monitor television (4), the The image is displayed. At this time, the light ben (5) is operated to register one frame's worth of image data, and at the same time, a window (8a) indicating the data reading range for recognition and its shift amount are set.

Next, when the first recognition command is issued while the image corresponding to the target workpiece (6) is displayed on the monitor television (4) in the recognition mode, the image processing circuit (3a) as shown in FIG.
Recognition processing is performed using the position and size of the window (8a) shown in a), and the results are displayed on the monitor television (4). Next, the image processing circuit (3a) responds to the second recognition command.
As shown in Figure 3(b), the recognition process is performed using the position and width of the window (8b), which is shifted to the left of the drawing by a specified pixel from the window (8a), and the result is Display on monitor TV (4). Hereafter, in the same way, window (8C) in Figure 3(C), window (8C) in Figure 3(C),
After repeatedly displaying the recognition processing results in the window (8d) of d), when a new window extends beyond the display frame, the process returns to the beginning. FIG. 4 is a flowchart showing these processing steps. Note that the window point in this case is always input as a variable called "WINDOW", and since each window does not have a point for reading an image, the point for one image is sufficient, and memory can be saved.

In the above embodiment, a light ben was used as the man-machine interface, but the above-described operations can also be performed using another human-powered device called a mouse.

In addition, in the above embodiment, the window is a monitor TV (4
) When the window exceeds the display frame, the recognition operation is repeated by returning to the beginning, but instead of this, the recognition operation may be repeated by sequentially shifting the window in the opposite direction.

〔Effect of the invention〕

As is clear from the above explanation, according to this invention,
During teaching mode, a window indicating the range of data reading for recognition and its shift 2 are set for the image displayed on the monitor TV, and during recognition mode, the image data displayed on the monitor TV and the storage device are set. Since the registered image data is compared with the registered image data while being shifted sequentially for each set window, it is possible to shorten recognition processing time and save memory capacity.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.
Figure 2 is a diagram showing the specific configuration of the same embodiment, and Figure 3 (
A) to (d) explain the operation of the same embodiment. FIG. 4 is a flowchart for explaining the operation of the embodiment, and FIG. 5 is a schematic configuration diagram of a conventional image recognition apparatus. (1); Television camera (2): Image input circuit (3a): Image processing circuit (4): Monitor television (5) Nilight Ben (9): Storage device (10): Registration means (11): Image recognition means , the same reference numerals in each figure indicate the same or corresponding parts. Agent Masuo Oiwa Diagram 3 Procedural Correction Power (Voluntary)

Claims (1)

[Claims] a recording device for storing image data of a recognition target; a television camera for photographing the recognition target; an image input circuit for performing binarization or other processing on the video signal of the television camera and outputting the image; A monitor television that inputs data processed by a circuit and displays an image to be recognized, and stores the image data displayed on the monitor television in the storage device during a teaching mode, and also controls the range of data reading in image recognition. a registration means for setting a window to be displayed and a shift amount thereof; and a registration means for setting a window to be displayed and a shift amount thereof; and a registration means for setting a window to be displayed and a shift amount thereof; 1. An image recognition apparatus comprising: recognition processing means for performing verification on each window and displaying the verification results on the monitor television.