JPH03259706A - Appearance inspecting device - Google Patents

Abstract

PURPOSE:To detect abnormality automatically and securely by counting picture elements perpendicular to a density gradient from an input image pickup image signal as to picture elements which are a specific value larger than the absolute value of a directional component having the density gradient in a constant direction, performing projection conversion, and comparing the evaluate value with a permissible error value. CONSTITUTION:A camera 2 is moved by a robot 1 to a specific position to picks up an image of an insulator 3 to be inspected and inputs its digital data to the image input means 4 of the appearance inspecting device 10 and a directional differentiating circuit 5 detects the density gradient of the insulator 3 in the fold direction to extract the directional component. As for the differentiated image, a projection data generating circuit 6 counts pictures parallel to the folds by positive and negative picture elements separately as to the picture elements whose absolute differential values are large than the specific threshold value. Projection data are regular when the insulator is normal, but irregular when not. The data are compared with the permissible error items stored in a storage means 9 and when an item is exceeded, an abnormal place is detected.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to an appearance inspection device for inspecting an abnormality in the appearance of an object to be inspected, such as an insulator, whose concentration repeats in one direction, for example, during a patrol inspection of a substation or the like using a mobile robot.

[Conventional technology]

FIG. 11 shows a circuit configuration diagram of a conventional image monitoring system disclosed in, for example, Japanese Unexamined Patent Publication No. 62-147888. an illuminance change calculation circuit that calculates the change in the illuminance meter 17; 19 is a histogram creation circuit;
20 is a continuous change determination circuit that determines continuous changes in a plurality of pixels; 2] is a changing object matching/identification circuit that identifies a changing object by comparing cutout data and projection calculation data of the changing object with a standard pattern 22 stored in advance; A circuit 23 is a reference image updating circuit for updating the reference image to the current image, and 24 is a reference image updating circuit for updating the reference image to the current image; This is a pixel change verification circuit that calculates and projects the cumulative number n for each line parallel to the coordinates. Next, the operation will be explained with reference to FIG. When attempting to detect abnormal changes, an image of the inspection area taken in advance is stored in the storage device as a reference image (Fig. 12(a)), and an image taken during the examination (the first
Perform differential processing with Figure 2 (b)) to detect an abnormal part (first
2 (C)). Next, detailed operations in FIG. 11 will be explained. In order to detect an abnormality using optical means, in the case of the conventional invention, normalized image comparison is performed with respect to changes in illuminance. In addition, a method is provided in which the images are compared again and the reference image of the background is updated without regarding the change in illumination as an abnormal change. In FIG. 11, three methods are employed for detecting signs of illuminance change, and each method is switched by an illuminance change detection method setting circuit 16. Among the three types, the first method calculates the change in the output of the illuminance meter 17 using the illuminance change calculating circuit 18, and the second method calculates the change in the output of the illuminance meter 17 by calculating the change in the illuminance change calculating circuit 18. The histogram is formed by a histogram creation circuit 19, and the third method is a method using a continuous change determination circuit 20 that determines continuous changes in a plurality of pixels. After making sufficient corrections due to illuminance changes in this manner, the current image and the reference image are compared and judged. That is, in the case of the first method, for example, when the amount of change in illuminance is larger than a set value, the reference image updating circuit 23 performs the determination by replacing the image with a reference image. In the case of the second method, a change in illuminance is detected using a histogram, a pixel change verification circuit 24 detects a pixel change, and a changed object comparison/identification circuit 21 compares it with a standard pattern 22 to make a determination. In addition, in the third method, when the continuous change determination circuit 20 determines that the number of continuous changes is greater than or equal to the set value R7, it is determined that the change is abnormal and an alarm is issued; If there is a change, it is determined that there is a sign of a change in illuminance, and a command to update the reference image is given to the reference image update circuit 23, and the reference image is changed to the current image and then detected.

[Problem to be solved by the invention]

Conventional image monitoring systems are configured as described above, so if there is a positional shift or change in size between the reference image and the inspection image, a normal part is likely to be detected incorrectly, and the difference between the two images There were issues such as the need for accurate positioning and size matching. This invention was made to solve the above-mentioned problems, and provides a visual inspection device that can detect only abnormal parts without requiring accurate positioning or size adjustment of the input image of the object to be detected. The purpose is to obtain.

[Means to solve the problem]

The appearance inspection apparatus according to the present invention includes a quantized image input means for capturing an image of an object to be inspected with a visual device and quantizing the image signal, and a density for determining each pixel of the image signal as a density gradient in a predetermined direction. a gradient detecting means, a counting means for counting and projecting the number of pixels in a direction perpendicular to the density gradient for pixels for which the absolute value of the direction component of the density gradient is equal to or greater than a certain value; an information processing means that determines an area where the count value is equal to or greater than a predetermined threshold value as an edge area and calculates an evaluation value, and compares the evaluation value with a tolerance value stored in advance in the storage means, and the evaluation value exceeds the tolerance error. and an abnormality detecting means for detecting the presence of an abnormal part, so that abnormalities in the object to be inspected are automatically detected.

[Effect]

The appearance inspection device according to the present invention captures an image with a visual device and counts the number of pixels in a direction perpendicular to the density gradient for pixels whose absolute value of a direction component having a density gradient in a certain direction is equal to or greater than a predetermined value from an input image signal. Projection conversion is performed, and the peak value of the projected image, the width of a predetermined position, the distance between peak values, and the evaluation value of the distance between positive and negative edges are obtained and compared with the allowable error value to detect abnormalities in the inspected object. The feature values are emphasized to reliably detect abnormalities even if there are changes in the position or size of the object in the image.

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic diagram showing an embodiment of the present invention. In the figure, 1 is a mobile robot, 2 is a camera, and 3 is an object to be inspected, such as an insulator. At a substation or power plant, a mobile robot 1 moves to an inspection location and automatically inspects the appearance of an insulator 3. Further, FIG. 2 is a system configuration diagram showing an embodiment of the present invention. In the figure, 4 is an image input means for receiving an output signal from the visual device 2, 5 is a directional differentiation circuit as a density gradient detection means for determining the density gradient in a predetermined direction, and 6 is the absolute value of the density gradient. a projection data generation circuit as a counting means for counting pixels having a value of a certain value or more in a direction perpendicular to the density gradient to obtain projection data; 7 is an abnormality detection means for detecting an abnormality in appearance from the characteristics of the projection data; 8; 1 is a processor serving as an information processing means for calculating an abnormality evaluation value and performing other operations; 9 is a storage means; and 10 is an appearance inspection device. Further, FIG. 3 is a flowchart showing the sequence of operations for inspecting the appearance. Roughly speaking, the mobile roof 1-1 images the insulator 3 to be inspected with the camera 2 after moving to a predetermined position. FIG. 4 is an example of an input image. The operation will now be described with reference to the flowchart shown in FIG. First, the mobile robot i moves to a predetermined position. Then, the insulator 3 to be inspected is imaged by the camera 2, and digital image input data is taken into the image input means 4 of the visual inspection apparatus 10 (step 5Tl). FIG. 4 is an example of an input image. The density gradient in the direction of the folds of the insulator 3 (from the top to the bottom in the figure) is detected from the human image, and the directional component is extracted. For this purpose, differentiation processing is performed using the directional differentiation circuit 5 having a certain directionality as shown in FIG. 5 (step 5T2). For example, if the folds are horizontal, a differential operator as shown in FIG. 6(C) is used. This differential operator only needs to be able to detect positive and negative direction components of edges such as folds. In the case of FIG. 6, the hatching in (a) indicates an image of the object, which is subjected to differentiation processing using, for example, the differential operator shown in FIG. 6(C). The concentration gradient on the three lines Xl, Yl, and Zl in Figure (a) is the same as the concentration gradient on the X2. Y2. This is an electrical signal from the directional differential circuit of Z2, where the sign represents the directionality and the concentration represents the strength of the concentration gradient. Therefore, the directional differentiation circuit 5 only needs to be an operator capable of detecting positive and negative direction components. The projection data generation circuit (counting means) 6 counts the number of pixels in the direction parallel to the folds, separately for positive and negative, for pixels whose absolute value of the differential value is greater than or equal to a predetermined threshold value for this differential image (step 5T3).
. Under normal conditions, regular projection data can be obtained as shown in Fig. 7, but if there is an abnormality as shown in Fig. 8, the projection data will also be obtained as shown in Fig. 9 (a) (b). )(c)
The regularity is lost as shown in (Step 5T4). Therefore, based on this projection data, the abnormality detection means 7 sets the area where the count value n is more than a predetermined threshold as the fold edge area, and the peak (value) coordinate PRj is set as shown in FIG.
(nRj+ yRj) + pt, t (nLJ
+yy), and the widths WRj, WLj of the projected image at a predetermined height are calculated by the processor 8 for the peak values yR2, yLj (step 5T5), and the distance between the peak coordinates 1)++i= (nRJ n++a++)+
DLJ-(n Lj n Lj.1) and the distance H4 between positive and negative edges are similarly determined by the processor 8, and (1
) is used as the evaluation value, and an abnormality is detected (step S
T6-5T9). Then, the respective allowable errors stored in advance in the storage means 9, for example, the distance between peak coordinates D, the peak value y, the width W of the projected image, the distance between positive and negative edges I] and the detected value of equation (1). Compare (Step 5TIO). For example, if all the evaluation values of equation (1) do not exceed a predetermined tolerance, it is evaluated as normal, and if they exceed the tolerance, it is detected as an abnormality (step 5T11). Further, an evaluation function such as a fuzzy function may be applied to these evaluation values to perform abnormality evaluation. In this embodiment, the inspection of insulators has been described, but the present invention can be similarly applied to objects such as stairs, ladders, and electric wires in which components in the same direction are repeated. 0

【Effect of the invention】

As described above, according to the present invention, there is provided an image input means for capturing an image of an object to be inspected using a visual device and quantizing an image signal, a concentration gradient detection means for calculating a concentration gradient in a certain direction from the image signal, and a concentration gradient detection means for determining a concentration gradient in a certain direction from the image signal. A counting means for projecting the number of pixels of the direction component of the gradient, an information processing means for calculating an evaluation value of a characteristic part from the projection-transformed data, and detecting an abnormal point by comparing the evaluation value with a tolerance error. Since the visual inspection apparatus is configured with the abnormality detection means, the features of the object to be inspected are emphasized and highly reliable inspection results can be obtained. In addition, since the inspection is performed using projection values, it is possible to easily deal with changes in size in the image, so there is no need for accurate positioning, and the mechanism is simple and inexpensive. Furthermore, since there are multiple evaluation criteria for abnormality determination, reliability is improved.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram showing an embodiment of an appearance inspection device according to the present invention, FIG. 2 is a block diagram showing a hardware configuration according to an embodiment of the invention, and FIG. An explanatory diagram showing the processing procedure, Fig. 4 is an explanatory diagram showing an example of an input image, Fig. 5 is an explanatory diagram of a directional differential image, and Figs. 6 (a), (b), and (C) are horizontal edge components detected. Figure 7 is a projection data diagram of the image in Figure 5 that counts the number of pixels in the horizontal direction. Figure 8 is an illustration of the input image at abnormal times. Figure 9 (a) (b) (c) is the 8th
Fig. 10 is an explanatory diagram of evaluation criteria for abnormality detection, Fig. 11 is a block diagram showing the hardware configuration of a conventional image monitoring system, and Fig. 12 is a conventional image monitoring system. It is a model diagram showing the principle of. In the figure, 2 is a camera (visual device), 4 is an image input means, 5 is a directional differentiation circuit (density gradient detection means), 6 is a projection data generation circuit (counting means), 7 is an abnormality detection means, 8
9 is a processor (information processing means), and 9 is a storage means. In addition, in the figures, the same reference numerals indicate the same or equivalent parts.

Claims (1)

[Claims] an image input means for capturing an image of an object to be inspected with a visual device and quantizing the image signal; a concentration gradient detection means for determining a concentration gradient in a fixed direction in each pixel of the image signal; a counting means for projectively converting pixels by counting the number of pixels in a direction perpendicular to the density gradient for pixels whose absolute values are above a certain value; an information processing means that determines an evaluation value for the edge region; and an abnormality detection means that compares the evaluation value with an allowable value stored in advance in a storage means and detects an abnormal location if the evaluation value exceeds the allowable error. Equipped with visual inspection equipment.