JPH09101374A - Method for inspecting wrong or missing externally mounted engine part - Google Patents

Abstract

PROBLEM TO BE SOLVED: To achieve a part inspection for checking, for example, the type, shape, and appearance of parts as well as a wrong or missing part by photographing an externally mounted engine part in a specific direction and performing the image processing of the picked-up image screen. SOLUTION: A CCD camera 2 picks up the image of an externally mounted part (for example, a hose) of an engine 5 in a specific direction, for example, from above and the side. An image processing device 3 divides an image pick-up screen vertically and horizontally into a plurality of small regions and performs gray image processing of the inside of the small region, thus measuring the brightness distribution within each region. Then, the presence or absence of the externally mounted part and the approximate shape can be detected. Then, the amount of change of the brightness distribution of each small region is detected for every small region. A personal computer 4 (incorporating a Fuzzy logic element) is used to compared detection data with reference data (acceptable article image pick-up screen), judge the agreement of the change amount with the reference data for every small region by the fuzzy inference, and inspect the wrong or missing externally mounted engine part according to the judgment over the entire small region.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an erroneous missing item inspection method for external parts attached to an engine.

[0002]

2. Description of the Related Art When manufacturing an automobile, after assembling an engine,
It is necessary to inspect whether or not the external parts, for example, harnesses or soft parts such as hoses are attached at predetermined regular positions. Therefore, in the case of an automatic machine, the arrow portions (Wa) (Wb) in the engine (1) shown in FIG.
The presence / absence check of parts at a predetermined position (Wc) is carried out by a contact sensor. Alternatively, it may be artificially checked visually.

[0003]

The problem to be solved by the present invention is that the conventional means for inspecting external parts for engine parts is different in the type and shape of parts, or changes in the mounting position due to replacement of workers, or a hose. The point is that it is not possible to perform an incorrect product inspection to check the appearance such as the routing shape of a class. Therefore, it is an object of the present invention to provide an erroneous deficiency inspection method capable of not only deficiency inspection of engine external parts such as hoses but also erroneous inspection of the type, shape and appearance.

[0004]

According to the present invention, an engine external component is imaged from a predetermined direction to divide an image pickup screen into a plurality of small areas, and each small area is subjected to grayscale image processing to measure a lightness distribution. Detecting the amount of change in the brightness distribution between each step and the small area, and comparing the detected data with the reference data to determine the degree of coincidence with the reference data of the above-mentioned amount of change for each small area by fuzzy inference, And a step of inspecting erroneous missing parts of external parts of the engine based on the judgment over all the small areas.

[0005]

1 (a) and 1 (b) show an embodiment of an erroneous missing item inspection method for external engine parts according to the present invention.
This will be described below with reference to (c) and FIGS. 2 (a) and 2 (b). First, FIG. 1A shows an apparatus configuration for carrying out an erroneous missing item inspection method for external engine parts according to the present invention. In the figure, (2) is a CCD camera, (3) is an image processing apparatus, and (4). ) Is a personal computer for fuzzy reasoning. CC above
The D camera (2) images external parts (for example, a hose) of the engine (5) from a predetermined direction, for example, from above and side,
The imaging screens (F) and (G) are shown in FIGS.
In addition, (F) is an image pickup screen when a hose non-defective product is attached, which can be selected as a reference image to be described later, and (G) is an image pickup screen when a hose is incorrectly installed.
e) is a wrong product with a defective mounting position.

The image processing apparatus (3) is shown in FIGS.
As shown in, the image pickup screens (F) and (G) by the CCD camera (2) are divided vertically and horizontally into a plurality of small areas (F1) (F2) ... (G1) (G2). Grayscale image processing (gray search) is performed on each of the small areas (F1) (F2) ... (G1) (G2) ... to measure the spatial lightness distribution within each area. Then, the lightness locally changes in the area depending on the presence or shape of an external component such as a hose and the like, so that it is possible to detect the presence or absence of the component and the general shape of the component from the lightness distribution in the small region. Then, the change amount (A1) of the lightness distribution between the small regions (F1) (F2) ... Or the change amount (Y1) of the lightness distribution between the small regions (G1) (G2) ... Detect every time. Therefore, when the non-defective imaging screen (F) is selected as the reference image, the variation amount (A of the brightness distribution between the detection data and the reference data {the small areas (F1) (F2) ... In the imaging screen (F) ...
1) ...} are compared, and for example, both are subtracted to calculate a difference (Q1 = Y1-A1) ... With reference data.

The personal computer (4) has a built-in fuzzy inference element and determines how much each variation (Y1) ... fits to the reference data (A1) ... for each small area (G1) (G2). Determined by fuzzy reasoning. For example, based on the difference (Q1) calculated by the image processing device (3), the change amount (Y1) for each small area (G1) (G2) is calculated by the centroid calculation of fuzzy inference.
The degree of coincidence (V1), which is how much the ... matches the reference data (A1), is determined. And the degree of coincidence (V1) ...
Is checked over the entire small area (G1) (G2) ... to inspect the type, shape and appearance of the external parts of the engine (5) to determine whether or not an erroneous product has occurred.

Next, the operation of the present invention based on the above configuration will be described. First, the CCD camera (2) images the external parts of the engine (5) from a predetermined direction (upper and side), and as shown in FIG. 1C, the image processing device (3) displays an imaging screen (G). Is divided into a plurality of small areas (G1). Therefore, the local lightness distribution in each area is measured for each small area (G1). Next, when attention is paid to each small area, for example, the small area (G8), adjacent small areas (G3) (G13) (G13)
7) Change in brightness distribution between (G9) and small area (G
8) and each brightness distribution difference (Y83) (Y813) (Y87) (Y89) between the surrounding small areas (G3) (G13) (G7) (G9) are detected. When fuzzy inference is performed by centroid calculation, the difference (Q83 = Y83-A8) between the detected data and preset standard data (A83) (A813) (A87) (A89)
3) Calculate ... Alternatively, the difference may be divided by the reference data to calculate the percent value of the difference for normalization. Therefore, the difference (Q83) ... is input to the register of the personal computer (4) for fuzzy inference, and the small area (G8) and (G8)
3) The degree of coincidence (V83) of each brightness distribution difference between (G13), (G7), and (G9) with the reference data is determined. Further, fuzzy judgment is performed over the entire small area (G1) ..., and the plural data are judged to check the presence / absence of external parts of the engine and the external shape, and inspect erroneous missing parts. At this time,
If the external components are imaged from two or more directions and the determination conditions are increased, the inspection accuracy is further improved.

Then, even if the shape and type of the external parts of the engine (5) change, the erroneous missing item inspection can be performed correspondingly. Further, when the surrounding light amount changes and the illumination condition changes, the lightness distribution itself changes, but the change amount of the lightness distribution between the small regions does not change. Therefore, the external parts of the engine (5) can be erroneously and deficiently inspected irrespective of changes in lighting conditions. Further, the number of divisions of the small areas (G1) ... May be increased or decreased as appropriate in consideration of the information processing time.

At this time, in the fuzzy inference algorithm, if the difference between the lightness distribution variation between the small regions and the reference data is small, the degree of coincidence becomes large and the imaged image of the engine external component is preset. If the difference approaches the reference image or the difference is large, the degree of coincidence becomes small and the reference image deviates. Therefore, the following fuzzy rules are created according to the above algorithm.

(I) IF Qn (n = 1 ...) = ZR (Zero), THEN
Vn (n = 1 ...) = PL (Positive Large) (II) IF Qn (n = 1 ...) = PS (Positive Small), THEN V
n (n = 1 ...) = PS (III) IF Qn (n = 1 ...) = PM (Positive Medium), THEN
Vn (n = 1 ...) = ZR Further, as a membership function for executing the fuzzy rule, a triangular membership function (Ma) (Mb) is set as shown in FIG. The membership function (Ma) is for the input part (% value), and the membership function (Mb) is for the output part (coincidence). Therefore, for example, if Qn = 5% as input data, the goodness of fit is 0.5 for rule (I), 0.5 for rule (II), and 0 otherwise. Therefore, output by the center of gravity calculation (coincidence)
(Vn) is around 2 and is quite close to the reference image. The above calculation is performed for all the small areas for the lightness distribution,
Inspect external parts of the engine (5) for erroneous missing parts.

Further, in the fuzzy reasoning, in addition to the judgment by the above-mentioned center of gravity calculation, there is a judgment means by probability. For example, FIG.
As shown in (b), the change amount (Y
1) Membership function (Mc) of fuzzy set of (where (ZRa) is reference data, (PSa) (NSa) is each fuzzy set of misalignment), and decision probability (Dn) is for each small region. Set to. Therefore, the adaptability (Ra) (Rb) with respect to the reference data and the positional deviation is detected from the change amount (Y1) of the lightness distribution between the small areas. And the goodness of fit (Ra)
The larger the value is and the smaller the matching degree (Rb) is, the closer to the reference data. Therefore, they are compared with the determination probability (Dn), and the matching degree (V1) ... Of the change amounts (Y1) ... Is determined. For example, Ra>
When Dn> Rb, the degree of coincidence (V1) is determined to be within the normal range, and the determination work is performed for each small area. Therefore, judgment is made from all judgment results over all small areas, for example, the number of times of normal judgment as a judgment standard, and an erroneous lack of external parts of the engine (5) is inspected.

[0013]

According to the present invention, the image pickup screen of the external component of the engine is divided into a plurality of small areas, and the lightness distribution is measured by performing grayscale image processing in each small area, and the lightness distribution between the respective small areas is measured. Since the degree of coincidence of the amount of change with the reference data is determined by fuzzy reasoning, and the erroneous missing parts of the engine external parts are inspected, not only the missing parts but also the erroneous parts can be inspected, and the illumination condition such as disturbance light fluctuation The accurate inspection can be performed regardless of the change of, the inspection is stable, and the accuracy is improved.

[Brief description of the drawings]

FIG. 1A is a device configuration diagram for carrying out an erroneous missing item inspection method for external engine parts according to the present invention.
(B) is a front view of a non-defective item imaging screen of an engine external component according to the present invention. (C) is a front view of an incorrect product imaging screen of an engine external component according to the present invention.

FIG. 2A is a waveform diagram of each membership function of the input / output unit for performing fuzzy inference in the method for inspecting erroneous missing parts of an engine external component according to the present invention. (B) is a waveform diagram of another membership function for implementing fuzzy inference of the method for inspecting erroneous missing parts of external engine parts according to the present invention.

FIG. 3 is a perspective view of an engine showing an example of a conventional engine external component.

[Explanation of symbols]

 2 CCD camera 3 Image processing device 4 Personal computer for fuzzy inference 5 Engine F, G Imaging screen F1 ..., G1 ... Small area

Claims (1)

[Claims] 1. A step of dividing an image pickup screen into a plurality of small areas by picking up an image of an engine-external component from a predetermined direction, performing grayscale image processing in each small area to measure a lightness distribution, and a step between each small area. The amount of change in the lightness distribution is detected, the detected data is compared with the reference data, and the degree of coincidence of the above-mentioned amount of change with respect to the reference data is judged by fuzzy inference for each small area, and judgment is made over all the small areas. And a step of inspecting erroneous missing parts of external engine parts.