JPH0216683A - Picture processor - Google Patents

Abstract

PURPOSE:To shorten recognizing time by providing a feature part to be the reference of a binarizing processing in a picture area to which a picture processing is to be executed and deciding optimum binarizing processing conditions with executing the picture processing on only the feature parts. CONSTITUTION:A TC camera 5 image-picks-up an object 2 optically, and the object 2 is recognized by executing a picture processing by scanning and a binarizing processing on a picture formed on the target of the TV camera 5. The picture processing is executed with using a feature part 9 provided separately from the object 2 at a fixed position in a picture area 8 in which the picture processing is to be executed. First, an objective picture is taken in, the small area involving the feature part 9 is cut off, and optimum binarizing conditions on the cut off small area, that is, a minimum scanning line number necessary for the recognizing of the object 2 is decided. Next, a binarizing processing is executed on the whole of the picture area 8 involving the object 2 based on the decided binarizing processing condition and the object 2 is recognized. Thus, recognizing time can be shortened.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image processing method for optically imaging and recognizing a target object.

(Prior art) An image processing method that optically images a target object, sequentially scans the entire image, and performs binarization processing to find and recognize the intersection with the target object. It is generally known (see, for example, Japanese Patent Laid-Open No. 61-30708).

In this case, in the image processing described above, an image obtained optically by a television camera is scanned in order to recognize the target object. This means that scanning will take longer. Therefore, a method has been adopted in which binarization processing is performed at a coarse scanning line interval at first, and then when the target object cannot be recognized, the scanning line interval is narrowed.

(Problem to be Solved by the Invention) However, in conventional image processing methods, in order to recognize a target object, it is necessary to perform binarization processing several times on the entire optically obtained image. Therefore, there is a problem in that it takes a long time to recognize the target object.

(Means for Solving the Problems) In view of the above, the present invention does not perform image processing for target object recognition from the beginning, but performs binarization separately from the target object in the area to be image processed. By setting a feature that serves as a reference for processing and first performing image processing on a narrow area that includes this feature, the optimal binarization processing conditions for image processing of the target are determined. Regarding image processing of a wide area including image processing, it is possible to recognize a target object by performing binarization processing once, thereby shortening the recognition time.

That is, the present invention is a method of optically capturing an image of a target object and performing image processing to recognize the target object. A characteristic part is provided as a standard for binarization processing,
Image processing is performed only on this feature to determine optimal binarization processing conditions, and based on these binarization processing conditions, the image area containing the target is binarized to recognize the target. It is characterized by this.

(Function) In the case of the above image processing method, image processing for only the characteristic part can be completed in a relatively short time because the area to be processed is small, and image processing for the image area including the target object can be completed in a relatively short time. Since this is performed based on the optimal binarization processing conditions obtained through the image processing of the characteristic portion, the target object can be recognized with one binarization process.

(Effects of the Invention) Therefore, according to the present invention, a feature part that serves as a reference is provided separately from the target object, the optimum binarization processing conditions are determined by image processing of a narrow area, and based on the optimum binarization processing conditions, the area including the target object is determined. Since the binarization process is performed, even if the area including the target object is large, the target object can be recognized with one binarization process, and the recognition time for this target object can be shortened. .

(Example) Embodiments of the present invention will be described below based on the drawings.

FIG. 1 shows an example in which a target object 2 placed on a pallet 1 is recognized by image processing and work is performed on it.

In the figure, 3 is a workbench on which the target object 2 is placed, 4 is a lighting device for the target object 2, 5 is a television camera that optically images the target object 2, and 6 is a working machine. In this case, by performing image processing on the image focused on the target of the television camera 5 through scanning and binarization processing,
The position, shape, or attitude of the target object 2 is recognized, and the controller 7 operates the work implement 6 based on this.

In the case of this example, the above image processing is carried out at a fixed position within the image area 8 (palette 1
A characteristic portion 9 provided separately from the target object 2 at a fixed position of
This is done using the . That is, this feature part 9 serves as a reference for the binarization process of the target object 2, and the feature part 9 is a level 2 that can be recognized.
The target object 2 can be recognized by the value conversion process.

The specific flow of image processing is shown in FIG.

First, an image of the object to be processed is captured, and the target object 2 is
In order to determine the binarization processing conditions, a narrow region including the characteristic portion 9 is cut out (steps S1 and S2). Then, the extracted narrow region is subjected to binarization processing at a relatively wide scanning line interval to extract the characteristic portion 9 (steps S3 and S4).

Next, the result of this extraction process is compared with preset teaching data, and the above-described cutting, binarization process, and characteristic part extraction are performed until the teaching data becomes equal to the teaching data (step Ss). In this case, the scanning line interval is gradually narrowed, thereby achieving the optimal binarization processing conditions for image processing of the target object 2, that is, the minimum number of scanning lines necessary for recognizing the target object 2. The number will be determined. Then, based on the determined binarization processing conditions, the entire image area 8 including the target object 2 is binarized to extract the features of the target object 2, and this target object 2 is recognized. (Step S6°Sy).

Therefore, the entire image area 8 only needs to be scanned once, and the previous scan for determining the binarization processing conditions can be performed in a narrow range at a predetermined position. , Overall, the time required to recognize the target object 2 is reduced compared to the case where image processing is performed while sequentially narrowing the scanning line sense for the entire image area 8 from the beginning. .

[Brief explanation of the drawing]

The drawings show an embodiment of the present invention, and FIG. 1 is an overall configuration diagram showing the relationship between the image processing means of the target object and the working machine, and FIG. 2 shows the relationship between the target object and characteristic parts in the image area. Floor plan, 3rd
It is a figure which shows the flow of drawing image processing. 2...Target, 5...TV camera,
8...Image area, 9...Characteristic part

Claims (1)

[Claims] (1) A method in which a target is optically imaged and image processing is performed to recognize the target, and the target is binarized separately from the target within the image area where image processing is to be performed. A feature part that serves as a reference for processing is provided, image processing is performed on only this feature part to determine optimal binarization processing conditions, and based on the binarization processing conditions, the image area including the target object is An image processing method characterized by performing binarization processing to recognize a target object.