JPS633382A - Image recognition device - Google Patents

Abstract

PURPOSE:To precisely judge the size of a reflective object by projecting light towards a camera through a space where the object is placed. CONSTITUTION:A light 3 irradiates the reflective object, a fish 1, carried by a steel-like conveyer, and the transmitting light is given to a scatter plate 4. Consequently, a camera 5 picks up the image of the fish 1, and no noise caused by reflection occurs. A picture signal from the camera 5 is stored in an image memory 7 through an A/D converter 6. A CPU 10 discriminates the size of the object, judging from the contents in the image memory 7 with the aid of a program in a memory 11. Thus the size of the reflective object can be precisely judged.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing device used for determining the size, area, length, perimeter, etc. of an object.

In particular, the present invention relates to a device whose target is an object that reflects light irradiated for imaging.

[Conventional technology]

The conventional device is disclosed in Japanese Unexamined Patent Publication No. 60-84682.

As in Japanese Patent Application Laid-Open No. 60-46378, a bandpass filter, a bandcapto filter, and a color filter are provided in front of the camera, and the size of the object is determined by emphasizing the image of the desired wavelength. .

[Problem that the invention seeks to solve]

In this case, if the object is an object that reflects the light emitted for imaging (for example, a fish), the reflected part will be black (and will be a bright color). There are cases where threshold settings cannot handle this problem, and in those cases, software is used to handle it.However, software processing is difficult, and noise remains.
Judgment of size etc. becomes inaccurate or unrecognizable.

An object of the present invention is to enable accurate determination of size, etc. when a reflective object is used as an object.

[Means for solving problems]

The above purpose is achieved by directing light toward the camera through the space in which the object is lost. [Operation] The camera images the shadow of the object. Therefore, noise due to reflection is eliminated and accurate judgment can be made.

〔Example〕

Hereinafter, the present invention will be explained in detail based on specific examples.

FIG. 1 shows a system configuration diagram in one embodiment of the present invention, and FIG. 2 shows a case with a diffuser plate (a) and a case without a diffuser plate (a).
The camera input in b) is shown.

First, in Fig. 1, l indicates the object to be recognized (in this example, a fish), 2 is a net-like conveyor that conveys the object, and 3 is a lamp that illuminates the object 1. , irradiates the object 1 from below to above. 4 is a semi-transparent diffuser plate, which is placed above the object l. The diffuser plate 4 is semi-transparent and has the property of blurring thin (small) shadows and making them invisible. For example, frosted glass.

5 is a camera installed above the diffuser plate 4;
An image within the field of view that includes the object is captured through the camera.

The A/D converter 6 converts an image signal (analog voltage signal) obtained by imaging an object with the camera 5 into a digital signal, that is, digital image data, and outputs the digital signal. A binarized signal is output using the threshold set in . Timing control section 8
The image signal from the A/D converter 6 is stored in the image memory 7 using the frog created. 9 is a bus for exchanging data. Central processing unit (hereinafter referred to as CPU) 1
0 reads the image data stored in the image memory 7 based on the image data processing program stored in the memory 11 and determines the size of the object l. The results are output to an external host computer or sorting machine via the input/output control@oka. Also, the keyboard 14
The data and commands set in are transferred to the CPU 10 via the input/output control section. The display unit 13 displays input data from the keyboard 14, the results of data processing by the CPU 10, the contents of the image memory 7, and the like.

Next, the operation of this configuration will be explained. If there is no diffuser plate 4, the net-like conveyor will also be imaged as shown in Figure 2 (Ca), but if the diffuser plate 4 is provided, the shadow part of the net will disappear as shown in (b), and only the shadow of the object will be captured. There's nothing left.

A recognition activation command is received through the input/output control unit, and recognition is started at that timing. The image signal from the camera 5 is binarized by the A/D converter 6 using the threshold value set on the keyboard 14, stored in the image memory 7, and then processed by the CPU 10.
The program in the memory 11 looks at the contents of the image memory 7, determines the size of the object (large, medium, small), and outputs the result to the outside via the input/output control section.

Therefore, the size of the objects on the conveyor can be easily and accurately sorted.

In the above embodiment, the conveyor 2 has a net shape, but plates may be arranged side by side at predetermined intervals, and the object may be supported by the end face of the cut plate in the thickness direction. That is, it is sufficient if the diffusion plate can eliminate the shadow of the thickness of the member supporting the object.

In the above embodiment, the conveyor is mesh-shaped because the object is conveyed by placing it on the upper surface of the conveyor, and a net-like member or the like is not required when the object is suspended and conveyed.

〔Effect of the invention〕

According to the present invention, because of the transmission method, there is no specular reflection of the object, so noise can be removed and the size etc. can be determined accurately.

[Brief explanation of the drawing]

Fig. 1 is a system configuration diagram showing one embodiment of the present invention;
The figure shows the image capture screen of the camera. l...Target, 2...Conveyor, 3.
-... Light, 4... Diffusion plate, 5...
...Camera, 6...A/D converter, 7...
...Image memory, 8...Timing control section, 9
...Pass, 10...Central processing section, 11
・・・・・・Input/output system Figure 11 A20 (Q) (G) Indication of the procedural amendment case Name of patent applicant (S+O+Tsuji Co., Ltd. Hitachi, Ltd. style Figure 2 of the drawing subject to human amendment Contents of the amendment Figure 2 of the drawing is amended as shown in the attached sheet. Figure 2 (Q) Cb)

Claims (1)

[Claims] 1. A light that illuminates the object, a camera that images the object, an A/D converter that digitizes the input signal of the camera, an image memory that stores the digitized image signal, and their calculations. What is claimed is: 1. An image recognition device equipped with an arithmetic processing circuit that irradiates light toward the camera through a space for placing an object in front of the camera.