JPS62229477A - Trouble diagnostic system - Google Patents

Abstract

PURPOSE:To omit a trouble whose defective state is to be judged by the human brain to improve the operability by reporting a trouble cause on a basis of relations among the trouble state recognized by a trouble recognizing means, the trouble state stored in a storage means, and the trouble cause. CONSTITUTION:If a defective image is read by an image reader 1, ranges of the white part and the black part of the defective image are identified and are sent as defective image data to a defective state recognizing device 2. Defective image data is sent to a preprocessing device 5, and noise and quantization error are corrected, and information compression is performed. Compressed defective image data is inputted to a variable density measuring device 6, a positioning device 7, and a comparing device 8. A feature extracting device 10 extracts features of the defective image existing in defective image data subjected to variable density and positioning corrections are extracted from comparison results of the comparing device 8. Feature patterns of typical indefective images stored in a defective image dictionary 11 are used on a basis of these extracted features to calculate the degree of coincidence with typical defective images by a defective image information identifying device 12. Such a data is sent to a trouble diagnostic device 3 as defective image information.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a fault diagnosis system, especially when the object of diagnosis is an image I.
AJ! I! The present invention relates to a failure diagnosis system for devices.

[Prior Art] In conventional fault diagnosis systems, an operator judges and inputs the answers to the questions of the fault diagnosis system using a computer terminal, which serves as diagnostic materials. Ta. However, when the diagnostic material is an image, the quality of the defective information sent to the failure diagnosis system is affected because the operator subjectively judges the defective image information that may exist in the image and returns the defective image information. It has the disadvantage that it is not constant.

[Problems to be Solved by the Invention] The present invention eliminates human judgment of defective states, equalizes the quality of information indicating defective states, and eliminates the trouble of human judgment of defective states. Provides a fault diagnosis system with improved operability.

Furthermore, the present invention provides a failure diagnosis system having a learning function that allows statistical data on failure causes to be collected and improves the statistical certainty of specialized knowledge for diagnosing failure causes.

[Means for Solving the Problem] As a means for solving this problem, the fault diagnosis system shown in FIG. and a result display device 4. The defective image recognition device includes a preprocessing device 5, a gradation measuring device 6, a positioning device 7, a comparison device 8,
The failure diagnosis device 3 includes a normal image dictionary 9, a feature extraction device 10, a defective image dictionary 11, and a defective image information identification device 12. It includes an inference device 14 consisting of an executor 18 and a stopper 19, and a failure state description device 15.

Further, the failure diagnosis system shown in FIG.
It includes U71, ROM72, and RAM73.

[Function] In this configuration, the defective image data from the image reading device 1 is sent to the preprocessing device 5 that constitutes the defective state recognition device 2, where it is corrected, etc. 8 is input. ? IA light measuring device 6
measures the density of the defective image, and the measurement results are sent to the comparison device 8 and defective image identification device 12. The alignment device 7 calculates the positional deviation between the defective image data and the normal images in the normal image dictionary 9, and the calculation results are sent to the comparison device 8 and the defective image identification device 12. The comparing device 8 corrects the defective image data based on the results of the gradation measuring device 6 and the alignment device 7, and compares the defective image data with normal images in the normal image dictionary 9. The feature extraction device 10 extracts the features of the defective image present in the defective image data that has been corrected for shading and alignment from the comparison result of the comparison device 8. Based on this feature, the defective image information identification device 12 calculates the degree of matching with the defective image using the characteristic pattern of the defective image stored in the defective image dictionary 11, and sends it as defective image information to the failure diagnosis device 3. It will be done. The failure diagnosis device 3 uses the failure image information stored in the failure state description device 15 to determine the cause of the failure when the condition part representing the failure state stored in the failure cause knowledge storage device 13 and the condition are met. The inference device 14 repeatedly outputs defective image information and failure cause knowledge as diagnostic information, and stops when there is no match. The diagnosis result display device 4 displays the diagnosis results obtained by the failure diagnosis device.

On the other hand, in FIG.
This system is replaced with an auxiliary RAM 73 for processing, and the CPU 71 performs the above operations.

[Embodiment] An example in which the present invention is implemented as a copying machine failure diagnosis system will be described as follows.

FIG. 1 shows the blocks of the fault diagnosis system and the information that passes between the processing work blocks. An image reading device 1 reads a defective image and sends the image data to a defective state identification device 2 .

The defective state identification device 2 recognizes defective image information based on the defective image dictionary and sends it to the failure diagnosis device 3. Failure diagnosis device 3
diagnoses the location of the failure from the defective image information based on the knowledge stored in the failure cause knowledge storage device and sends it to the diagnosis fruition device 4 as diagnostic information.

The diagnosis result display device 4 displays the diagnosis results on a display device.

Furthermore, details of each block will be explained below with reference to FIGS. 2 and 3. FIG. 2 is a block diagram of the defect status recognition device 2. FIG. 3 is a block diagram of the failure diagnosis device 3.

The image reading device fW1 is a device that identifies white parts and black parts of a defective image and captures them into the defective state recognition device 2. The defective state recognition device 2 uses the image data sent from the image reading device 1 and the normal image dictionary 9 and the defective image dictionary 11 to
This is a device that recognizes defective image information and sends it to the failure diagnosis device 3, and is composed of a preprocessing device 5, a gradation measuring device 6, an alignment device 7, a comparison device 8, a feature extraction device 10, and a defective image information identification device 12. It is a device that can be used. The failure diagnosis device 3 is
The inference device 14 uses the failure cause knowledge storage device 13 that stores knowledge for diagnosing the cause of failure obtained from human experts from the defective image information device sent from the defect status recognition device 2.
This device performs failure diagnosis by detecting failure conditions that are not yet known, and sends diagnostic information to the diagnosis result display device 4.

The operation of the fault diagnosis system will be explained according to FIGS. 2 and 3. When a defective image is read by the image reading device 1, the ranges of white and black portions of the defective image are identified and sent to the defect state recognition device 2 as defective image data. What is a bad image here?
This is an image of a test pattern copied by a copying machine that is subject to failure diagnosis. The defective image data is sent to the preprocessing device 5 that constitutes the defective state recognition device 2, and is sent to the image reading device a1.
Noise and quantization error caused by i are corrected, sampling processing is performed, and information is compressed. The compressed defective image data is manually input to a gradation measuring device 6, a positioning device 7, and a comparing device 8. The density measurement device 6 measures the density of the white background and black background of the defective image, and the measurement results are sent to the comparison device 8 and the defective image identification device 12. The alignment device 7 calculates the positional deviation between the defective image data and the normal images in the normal image dictionary, and the calculation results are sent to the comparison device 8 and the defective image identification device 12. The comparing device 8 corrects the defective image data based on the results of the gradation measuring device 6 and the alignment device 7, and compares the defective image data with normal images in the normal image dictionary 9. Here, the normal image is information obtained by reading data obtained by copying a test pattern with the image reading device 1 and processing it in the preprocessing device 5 in a state where there is no failure. Is the feature extraction device 10 based on the comparison result of the comparison device 8? The features of the defective image present in the defective image data that has been corrected for IA light and zero alignment are extracted. Based on these characteristics, the defective image information identification device 12 calculates the degree of matching with the typical defective image using the characteristic pattern of the typical defective image stored in the defective image dictionary 11. This data is sent to the failure diagnosis device 3 as defective image information. The defective image dictionary 11 stores coded defective image data shown in FIG. 4, and FIG. 5 shows an example of defective image information identified by the defective image identification device 12.

As shown in FIG. 3, the failure diagnosis device 3 includes a failure cause knowledge storage device 13, an inference device 14, and a failure state description device 15.
It consists of The failure cause knowledge storage device 13 has a sixth
As shown in the figure, fragmentary knowledge of failure causes is stored in the form of rules, consisting of a condition part representing a defective state and an execution part representing the cause of failure when it can be inferred that the condition is met. The defective state description device 15 is the inference device 14
This is a device that temporarily stores defective image information that can be rewritten by. The inference device 14 is a device that repeatedly outputs the defective image information stored in the defective state description device 15 and failure cause knowledge as diagnostic information, and stops when no matching information is found.

Details of the inference device 14 will be explained using FIG. 3. The inference device 14 includes a comparator 16, a rule selector 17, and an executor 18.
, a stopper 17. The comparator 16 compares the contents of the failure state description device 15 and the failure cause knowledge storage device 13, and outputs all rules in which the condition part of the failure cause knowledge matches each piece of information in the failure image information. If there is no matching rule, a signal is sent to the stopper 19 and the inference device 14 is stopped. If a matching rule exists, that rule is sent to the rule selector 17.

The rule selector F selects from among the rules from the comparator 16,
The number of matches of the conditions in the comparator 16 is checked, and the rule with the maximum -politics/religion is selected. - If there are multiple rules with the largest political religion, the one with the lowest rule number is selected. The execution part of the selected rule is sent to the executor 18 and output as diagnostic information. The diagnosis result is sent to the defect status description device 15.
is accumulated, and the comparator 16 is activated again. Note that this executor 18 writes the execution part of the rule selected by the rule selector 17, p! Can be replaced. This is a function to count the frequency of failure causes and obtain reliable knowledge of failure causes. The diagnosis result display device 4 displays the diagnosis results obtained by the failure diagnosis device to a person using the copying machine failure diagnosis system.

FIG. 7 shows a block diagram of a fault diagnosis system according to another embodiment. The image reading device 1, defect/normal image dictionary 9, defect image dictionary 11, failure cause knowledge storage device 13, and diagnosis result display device 4 shown in FIGS. This is a system in which the CPU 71, a ROM 72 in which a processing program is stored, and a RAM 73 for assisting processing.

This will be explained according to the diagnostic flowchart shown in FIG. First, in step S81, defective image data is read from the image reading device, and in step S82, the above-mentioned preprocessing is performed on the defective image data. In step S83? It is checked whether there is a defect in Qt'A, and if there is no defect, the process jumps to step S87. If there is a defect, the shading defect is stored as part of the defect information in step S85, and the shading of the preprocessed defective image data is corrected in step S86. Step S8
In step S7, it is checked whether there is a defect in the position of the image, and if there is no defect, the process jumps to step S91.

If there is a defect, the positional defect is stored as part of the defect information in step S89, and the position of the defective image data is corrected in step S90. The defective image data at this point is data that does not include defects in shading and defects in position. This defective image data is compared with the normal image data of the normal image data 9 in step S91. If it is the same as normal image data,
The process jumps to step S96. If not normal, step S
In step S93, the defective features are extracted, and in step S94, the defective image dictionary 11 is searched based on the extracted features, and in step S9
In step 5, the search results are stored as remaining defective information. At this point, the creation of defective image information consisting of shading defects, position defects, and other defects is completed.

Next, in step S96, it is checked whether there is defective image information, and if there is no defective image information, the diagnosis is ended. Steps 397 to 103 are repeated while any defective image information exists.

In step S97, the failure cause storage device 13 is searched based on the defective image information, and all information regarding the cause is read out. In step S98, the cause with the largest number of appearances is extracted. If one cause is determined, the process proceeds from step S99 to step 5IOI. If there are two or more causes, the cause with the highest priority is selected in step 5100. Step 5 The number of occurrences of the cause selected in IOI is counted up, the priority is increased for each piece of defect information, and in step 5102, the cause is output to the diagnostic result display device 4, and the corresponding defect information is reset. , enter the next diagnosis.

In addition, the priority may be stored in the storage location in the failure cause storage device 13, or the priority or the number of occurrences may be stored in each data. When considering application to normal image definition 1! F9, defective image dictionary f1
1. This can be easily realized by exchanging the contents of the failure cause knowledge storage device 13 with data for each laser beam printer and facsimile. As explained above, by attaching an image reading device and a fault condition recognition device to a fault diagnosis device, it is possible to indicate a fault state without ambiguity compared to a method in which a human determines the fault state of the image and inputs it. This has the effect of equalizing the quality of information, enabling more accurate fault diagnosis, and improving operability. Furthermore, when determining failure cause knowledge, collecting statistical data on the failure cause obtained using the image reading device and the failure state recognition device has the effect of obtaining more reliable failure cause knowledge.

[Effects of the Invention] According to the present invention, human judgment of defective conditions is eliminated,
It is possible to provide a failure diagnosis system that achieves uniform quality of information indicating a defective state, eliminates the trouble of human judgment of a defective state, and improves operability.

Furthermore, it is possible to collect statistical data on the causes of failures, and it is possible to provide a failure diagnosis system with a learning function that improves the statistical certainty of specialized knowledge for diagnosing the causes of failures.

[Brief explanation of drawings]

Fig. 1 is a block diagram of the fault diagnosis system of the embodiment, Fig. 2 is a block diagram of the fault state recognition device, Fig. 3 is a block diagram of the fault diagnosis device, Fig. 4 is an example of the contents of the fault image dictionary, Figure 5 is a side entrance for defective image information, Figure 6 is a side entrance for failure cause knowledge, Figure 7 is a block diagram of the other side's failure diagnosis system, and Figure 8 is an operation flowchart of the failure diagnosis system in Figure 7. be. In the figure, 1... Image reading device, 2... Defective image recognition device, 3... Failure diagnosis device, 4... Diagnosis result display device, 5... Preprocessing device, 6... Shading Measuring device, 7
... Alignment device, 8. Comparison device, 9. Normal image dictionary, 10. Feature extraction device, 11. Defective image dictionary, 12. Defective image information identification device, 13.・
・Failure cause knowledge storage device, '14... Reasoning device, 15
...Bad state description device, 16...Comparator, 17...
- Rule selector, 18... Executor, 19... Stopper, 71... CPU. 72...ROM, 73...RAM. Patent applicant Canon Corporation Figure 1 Figure 2 Figure 3r11J Analysis information board Figure 4 Figure 5 Figure 6

Claims (10)

[Claims] (1) In a fault diagnosis system for diagnosing a fault in an image processing device, there is an image data input means for inputting image data, and a fault in which the fault state of the image processing device is recognized from the defective image data input by the image data input means. a recognition means; a storage means for storing a failure state and a cause of failure in association with each other; What is claimed is: 1. A failure diagnosis system comprising a cause reporting means for notifying a cause, and diagnosing a failure of an image processing apparatus based on image data. (2) The failure recognition means includes a feature extraction means for extracting features of the defective image data by identifying defects in shading, position, and defects other than the shading and position of the image processing device from the defective image data. The failure diagnosis system according to claim 1, characterized in that: (3) The failure recognition means is characterized by comprising feature extraction means for extracting features of the defective image data by comparing the defective image data corrected for defective shading and position with normal image data stored in advance. A failure diagnosis system according to claim 2. (4) The failure diagnosis system according to claim 3, wherein the failure recognition means recognizes the failure state by comparing the characteristics of a defective image stored in advance with the characteristics of the defective image data. (5) The cause notification means reads out from the storage means all the causes of failure associated with the failure state recognized by the failure recognition means, and notifies the cause of failure that appears most frequently. 4. The failure diagnosis system according to item 4. (6) In a fault diagnosis system for diagnosing a fault in an image processing device, an image data input means for inputting image data, and a fault in recognizing a fault state of the image processing device from the defective image data input by the image data input means. a recognition means; a storage means for storing a failure state and a cause of failure in association with each other; A cause informing means for notifying the cause, and a changing means for changing the stored contents of the storage means according to the failure state and the number of occurrences of the cause of the failure, and performs a failure diagnosis of the image processing apparatus based on the image data. A fault diagnosis system characterized by having a function of improving fault diagnosis. (7) The failure recognition means includes feature extraction means for extracting features of the defective image data by identifying defects in shading, position, and defects other than the shading and position of the image processing device from the defective image data. The failure diagnosis system according to claim 6, characterized in that: (8) The failure recognition means is characterized by comprising feature extraction means for extracting features of the defective image data by comparing the defective image data corrected for defective shading and position with normal image data stored in advance. A failure diagnosis system according to claim 6. (9) The failure diagnosis system according to claim 6, wherein the failure recognition means recognizes the failure state by comparing the characteristics of a defective image stored in advance with the characteristics of the defective image data. (10) The cause notification means reads out from the storage means all the causes of failure associated with the failure state recognized by the failure recognition means, and notifies the cause of failure that appears most frequently. 10. The failure diagnosis system according to item 9.