Patent Grant
6862414
1. Field of Invention
This invention relates to malfunction diagnosis of banding defects in document processing systems based on defect feature analysis and machine data analysis.
2. Description of Related Art
In document processing systems, it is well known that customer satisfaction can be improved and maintenance costs reduced if problems with copiers and printers can be fixed before they become serious enough to warrant a service call by the customer. Systems exist that enable printers and copiers to call for service automatically, when sensors detect certain operating parameters outside of permissible ranges. Generally, these systems take effect after a threshold has been reached where the degradation in image quality is directly observable by the user. Given the large number of operating parameters that need to be tracked during operation, a specific defect at a certain level may or may not be a significant problem. The overall affect of a specific defect depends on the cumulative values of the other parameters in the system. Systems do exist that attempt to diagnose failures in document processing systems based on image quality analysis of the print. Such a system for the diagnosis of copier performance over telephone lines is described in U.S. Pat. No. 5,365,310, incorporated herein by reference in its entirety.
A system and method for automatically diagnosing image quality defects is described in U.S. patent application Ser. No. 464,596, filed on Dec. 16, 1999, now U.S. Pat. No. 6,665,425 and commonly owned with the subject invention. The disclosure of this application is incorporated herein in its entirety. In this system, image quality problems occurring in a document processing system are identified by applying image processing and pattern recognition techniques to analyze specific test patterns. The results are analyzed in conjunction with known generic or experienced unit specific machine data in a diagnostic expert system to determine the cause of the problem. Further systems are provided to determine and execute a recommended service procedure, such as a user guided repair; scheduled maintenance service, parts replacement, and the like.
One of the most frequently encountered image defects is the banding defect which is symptomatic of a wide variety of malfunctions in the performance of a document processor. These defects manifest themselves as one or more lines, streaks, or bands extending across the document copy parallel to the short edge or the long edge, and may be periodic or non-periodic. They are caused by worn or damaged parts, foreign matter, electrical malfunctions, misalignment and other sources, within the print engine and the associated feed mechanisms and controls.
It is a purpose of the methods and systems described in this application to provide a means by which banding defects may be identified, analyzed, and quantified. It is a further purpose to diagnose a probable source of the defect and propose to the user an appropriate repair or correction.
The systems and methods of this invention provide automated diagnosis and repair of malfunctions in document processing systems based on a banding defect analysis.
The banding defect analysis is accomplished by using a series of computer modules and algorithms to analyze specific test patterns via techniques such as image processing and pattern recognition. As a result of analyzing the output of the document processor in response to specifically designed test patterns, the banding defects are characterized in terms of parameters that quantify the overall impairment of perceived image quality caused by the defect, as well as additional quantitative parameters that characterize the nature of the defect. The banding defect analysis is based on image analysis techniques including fourier analysis, filtering and human visual perception modeling.
Key features are extracted from the banding defect analysis output to provide data for determining possible sources of the defect. A series of algorithms for extracting pre-determined features of the banding defects from the outputs of the banding defect analysis is stored. These algorithms utilize threshold values for such features and device specific specifications. Using these algorithms and the outputs of the banding defect analysis, the key features of a particular banding defect are established.
In order to diagnose the source of the defect the key features are subject to analysis in a diagnostic engine, such as a bayesian network, a neural network, or a model based or rule based, diagnostic engine to determine the probable source of the defect. The diagnostic engine utilizes a matrix of malfunctions and defect features symptomatic of a particular malfunction. Once a source malfunction is identified within a predetermined certainty by the diagnostic engine, it is submitted to a repair planning module which determines the recommended service. The repair module refers to a stored reference file or table which relates the identified source to the repair service.
The diagnostic system of this application, includes an image quality analysis module which identifies and characterizes a banding defect in terms of quantitative parameters and generates key features of the banding defect for further analysis. The output of the image quality analysis module is preprocessed for use in a diagnostic engine. The data is processed in diagnostic engine to correlate the key features of the banding defect to a malfunction which is the possible source of the defect. A recommended repair or service is selected by the repair planning module. The results are presented to the user through a user interface. A diagnostic controller controls and coordinates the operation of all of the modules. A memory is provided in operative association with the processing components of the diagnostic system to store the algorithms and data used in the analysis and diagnosis.
The preferred embodiments of the invention will be described in detail, with reference to the following drawing in which:
FIG. 1(a) is a functional block diagram illustrating an embodiment of a diagnostic system, according to the system of this application;
FIG. 1(b) is a functional block diagram of a document processor system which is adapted for use with the diagnostic system of FIG. 1(a);
As used in this application, document processing systems, include analog and digital copiers, printers, scanners, facsimiles, and multifunction machines. Said systems include those based on all direct and indirect marking technologies, both color and black and white, such as xerography, ink jet, liquid ink, lithography, and the like.
FIG. 1(a) illustrates an embodiment of a general purpose diagnostic system 10, adapted, according to the system of this application, to the analysis of banding defects. Banding defects are the most common image defect experienced in document processing systems and are the most frequent cause of service calls. The analysis of the lines, streaks, and bands which comprise banding defects is no easy task, as such defects can reflect some 30 to 40 malfunctions which may occur in a document processing system.
The diagnostic system 10 operates in association with a document processing system 100. The diagnostic system 10 can be part of a document processor, multifunction machine, printer, etc., or could be part of a general purpose computer server connected to the machine, or could be implemented as a stand alone appliance having appropriate plug in capability for operation with a variety of machines in many different environments. For illustration, the basic components of document processing system 100 are shown in FIG. 1(b) and consist of a print engine 110 which is served by a document feed 120 and a scanner 130. Scanner 130 could also be part of the diagnostic system 10, as shown in FIG. 1(a). In some instances, such as a stand alone appliance or where there is no scanner available in the document processing system, a separate diagnostic scanner could be used in addition to or instead of the scanner of the document processing system. A system controller 140 provides operating control of the system 100 in conjunction with memory 150. An array of sensors 160 is distributed throughout the system to monitor the performance of system 100 at key points. The sensors generate current system data which can be stored in memory 150 to provide historical and status data to assist in analysis of defects. Further system performance data can be obtained by monitoring operating signals and other characteristics of the document processing system 100. For simplicity such monitoring function is encompassed in the sensor array module for the purposes of this application.
As indicated above, the document processing system 100 can include a wide variety of components and architectures. The embodiment of FIG. 1(a) is presented only as an example for illustrating the operation of the banding defect analysis system of this application. The system of this application is intended for use with document processing systems in general.
The diagnostic system 10, as shown in FIG. 1(a), includes an image quality analysis module 1 which identifies and characterizes a banding defect in terms of quantitative parameters and generates key features of the banding defect for further analysis. Additionally, the user is prompted to input additional features describing the defect, such as for example, by the selection of one of a set of icons or images, or by answering a set of specific questions. The output of the image quality analysis module 1 and the user input data is adapted for use in a diagnostic engine 2 by a preprocessor 3. The data is processed in diagnostic engine 2 to correlate the key features of the banding defect to a malfunction which is a possible source of the defect. A probability of causation is be evaluated and a recommended repair or service is selected by the repair planning module 5. The results are presented to the user through user interface 6. User interface 6 may include a display screen and appropriate keypad (not shown).
The diagnostic controller 4 controls and coordinates the operation of all of the modules 1 through 7 and 21-23. A memory 9 is provided in operative association with the processing components of the diagnostic system 10 to store the algorithms and data used in the analysis and diagnosis. Memory 9 may also be adapted to track the operation of the diagnostic system 10, by logging and categorizing data. In this manner a historic data base of error correction may be maintained for future reference by diagnostic engine 2.
A test pattern selection module, as shown in FIG. 1(a), selects the appropriate test patterns to be copied or printed and subsequently scanned and the sequence in which the selected test patterns are to be copied or printed and scanned and analyzed. The sequence may be predetermined and stored as a look-up table, or, can be determined dynamically, during the diagnostic process based on optimizing criteria such as minimal number of sheets to scan, maximal discriminating power of the tests, or maximal accuracy of diagnosis. In addition, in specifying the test patterns and their sequence, the test pattern selection module may select the specific region of the scanned image to be analyzed, and the set of features that need to be extracted from the results of the banding defect analysis.
The diagnostic system 10, as shown in FIG. 1(a), is adapted to consider all of the data generated by the image quality analysis module 1 and eventually, using historical and experimental data relating to the causes of banding defects and data relating to the service fixes for such causes, present instructions to the user to accomplish a recommended service agenda.
At the end of a diagnostic cycle, after an automated repair system or the user have performed the appropriate repair action, the correction verification module 23 prompts the user or the system to run additional tests or print additional test prints to determine if the correction that was accomplished actually was successful in eliminating the banding defect. If not the user may be directed to perform additional repair actions or to contact service personnel.
In the alternate embodiment shown in
As part of the alternate embodiment, shown in
The components of the document processing system 100 are operationally interconnected by means of a bus system 170 and may be connected to the diagnostic system 10 by means of a connecting network 8. The components are shown for clarity as independent functional modules, but it should be noted that the related functions can be implemented by a central microprocessor with algorithms controlling the functions. In addition it should be appreciated that any one of, or a portion of the components of the diagnostic system can be located anywhere, including on the actual document processing system itself, on a distributed network, or an adjacent or remote diagnostics location. The flexibility of location could be implemented through the use of wired or wireless links or any other known or later developed element(s) that is capable of supplying electronic data to and from the connected elements. Furthermore, network 8 can be any one of, or combination of, a bus system, a direct serial connection, direct parallel connection, a distributed network such as an intranet, a local area network, a metropolitan area network, a wide area network, a satellite communication network, an infrared communication network, the Internet, or the like. Furthermore the diagnostic system 10 and the document processing system 100 may not be connected by any electronic means at all in particular in the embodiment of
As shown in
Banding defect analyzer 13 is a computer module capable of executing image analysis algorithms available in memory 9. Banding defect analyzer 13 processes the image data to characterize the banding defect in terms of quantitative parameters. The image quality analysis module 1 analyzes the image, using commonly known image processing techniques such as, for example, Fourier transform analysis, band-pass filtering, histogramming, edge detection, 1-D projections, segmentation, classification, artifact detection, FIR filtering, wavelet analysis, statistical analysis, pattern recognition techniques, or the like, to evaluate image quality parameters and/or identify defects in the image(s). A typical banding defect analysis of the image data will provide data such as the following:
It is to be noted that other signals related to visual perception such as reflectance values, or CIE XYZ values, or physical measurements such as reflectance values measured through scanner RGB filters may be used in place of the L* signal.
After the banding defect analysis is performed on a scanned image, further processing in feature extraction processor 14, determines/generates the key features of the banding defect based on predetermined defect characteristics stored in processor 14. These characteristics or features will be selected from a group, such as the following:
The extraction process is accomplished by executing a series of algorithms based on the banding defect analysis outputs. The basic steps of the analysis involve the comparison of the banding analysis outputs or quantities derived therefrom, to predetermined maximum or minimum thresholds and device specific specifications. Examples of device specific specifications include the distance between fuser fingers, diameter of the donor roll, and length of the photoreceptor pitch, in the document processing system whose malfunctions are to be detected. The extraction process algorithms are described in simple terms below to illustrate the analysis.
(i) Defect Presence: Check if the VBS number is greater than a given threshold. This threshold can be obtained from the IQ specifications for the copier/printer to be diagnosed, or by analyzing sample prints obtained from a copier/printer with no IQ related failures and obtaining the average VBS number over a set of prints. This algorithm can be used to check for the presence of the defect in images obtained by using different modes on the document processing system such as making a copy using the platen glass, making a copy using the automated document feeder, printing an internal test pattern, or printing an image submitted through the digital front end.
(ii) Defect Orientation: If the VBS number is greater than the threshold value for a segment of the scanned image that runs along the long edge of the sample print, then defect orientation=short edge; if the VBS number is greater than the threshold value for a segment of the scanned image that runs along the short edge of the sample print, then defect orientation=long edge.
(iii) Defect Type: Find the peak values within the L* peaks, or the visually filtered peaks, or, both (LSMinValue and LSMaxValue, or VFMinValue and VFMaxValue, or both) that exceed given thresholds. These thresholds can be predetermined fixed numbers, or can be based on the current image and can be equal to a significant statistic such as mean+3 std deviation, where the mean and standard deviation are computed over all Min and Max peak values across the image. Peak values that exceed the threshold will be characterized as “isolated defects”. Further, depending on whether the peak is a negative(Minvalue) or a positive one (MaxValue), the Defect Type will be set to “IsolatedDark” or IsolatedLight”. All peak values that do not cross the threshold to be marked as isolated defects but which still exceed a lower threshold will be marked as “potential defects”. If the number of negative potential defects is greater than a specified threshold, then Defect Type=“Uniform Dark”; If the number of positive potential defects is greater than a specified threshold, then Defect Type=“Uniform Light”;
(v) Defect Separation: Compute the distance between defects marked as isolated or potential by computing the difference between their locations
(vi) Defect Spectral Properties: Check if the Fourier transform amplitudes at specific frequencies are significantly higher than average; check if the Fourier transform amplitudes are significantly lower at high frequencies.
(vii) Defect Change with Change of Gray level: Check if defect is present (as explained in (i) above) in a light image and not in a dark image and vice versa; Check if the defect polarity (as explained in (iv)above) changes from light to dark or vice versa in images of differing gray levels.
(viii) Defect Enlargement/Reduction with Image: Compare the size of the defect in two images, one obtained by copying an original image (on the machine to be diagnosed) at one magnification level and the second obtained by copying the original at a higher/lower level of magnification. If the defect size is greater (lesser) with enlargement (reduction), defect is said to enlarge (reduce) with image.
The extracted features are converted to a format acceptable to the particular diagnostic engine used, for example event format, in diagnostic preprocessor 3 and submitted to the diagnostic engine 2 for additional analysis. Diagnostic engine 2 is a computer module capable of executing algorithms which apply reasoning techniques such as qualitative reasoning, probabilistic reasoning, or fuzzy reasoning, and could be rule based, object based, or model based, or case based. A bayesian network is used in the preferred embodiment, but it should be noted that the system could be adapted to other forms of diagnostic algorithms. The output of the diagnostic engine 2 is sent to the diagnostic controller 4 for correlation with a repair scheme from repair planning module 5 and presentation to the user via user interface 6.
In order to provide the diagnostic engine 2 with data from which to correlate particular extracted features with a particular malfunction which may cause the particular defect, a matrix of malfunctions is compiled with the likely defects that would result. For example in a particular document processing system the following malfunctions may occur which might result in a banding defect:
To isolate the resulting banding defect, a test pattern having a uniform gray image, at a certain area coverage is scanned using several variations of steps. Since some of the defects caused by one or more of the above malfunctions only appear in alternate copies or only when the document feeder is used, the user is directed to make two sequential copies of the selected test pattern and also to make copies with and without the document feeder.
A damaged fuser finger, for example, will cause repetitive streaks separated by a specific distance which correlates to the distance between the fuser fingers. Therefore, in analyzing the defect separation feature, this malfunction can be readily identified or eliminated depending on the value of defect separation. If such a malfunction is one of the set of malfunctions that the diagnostic system is designed to detect, then separation between defects would be extracted as a key feature.
A matrix of such defect data based on experience with a fleet of document processors or with a particular processor can be compiled for use by the diagnostic engine 2. An example of such a matrix is shown in the table of FIG. 8. When the diagnostic engine used is a probabilistic one, such as a bayesian engine, the table includes conditional probabilities of the defect features given the malfunctions, as well as a prior probabilities of the malfunctions. An example of a defect feature matrix with probabilities is shown in FIG. 9.
The diagnostic engine 2 performs the task of reasoning with the multiple sources of information to isolate the cause of the malfunction indicated by the banding defect. This can be an iterative process, for example, test pattern selection module 21 may decide, based on an initial diagnosis, that additional test patterns are to be scanned. In addition certain specific image quality parameters may be emphasized for evaluation and specific defect features may be extracted in order to further isolate the malfunction.
In the embodiment of FIG. 1(a) Test pattern selection module 21 may be based on a simple look-up table that maps malfunctions to specific test patterns or it may be based on an optimization scheme. When the test pattern selection module is implemented as a simple lookup table, at any time during the iterative diagnostic process, the malfunction that appears most probable at that point may be chosen and a test pattern associated with that malfunction will be chosen as the next test pattern to scan and analyze. Where there is more than one test pattern associated with a failure, the test patterns will be scanned and analyzed in the order in which they are listed in the table, with the constraint that any test pattern scanned and analyzed already will not be repeated. In a case where the test pattern selection module is based on an optimization scheme, the next test pattern chosen will be that particular test pattern among the given set of test patterns that maximizes an optimization criteria, for example: the test pattern that maximizes the entropy function, provides the best discriminating information to isolate the malfunction, minimizes the time taken to reach a diagnosis, or minimizes the overall cost of running the diagnostic system, or a combination of the above.
Based on the test pattern chosen by the test pattern selection module, the document processing system 100 is directed to produce one or more print samples of an original test pattern stored either in its internal memory, or on a network drive in a distributed network environment. The print samples are then forwarded to the scanner 130 or the external scanner for digitizing. Alternately, the customer, the customer service engineer, or the like, can make copies of hard copy original test patterns and then scan them.
Alternately, in the embodiment of
As in the case of the test pattern selection module 21, the test selection module 22 may be based on a simple look-up table that maps malfunctions to specific tests or it may be based on an optimization scheme. When the test selection module is implemented as a simple lookup table, at any time during the iterative diagnostic process, the malfunction that appears most probable at that point may be chosen and a test associated with that malfunction will be chosen as the next test to perform. Where there is more than one test associated with a failure, the test will be performed in the order in which they are listed in the table, with the constraint that any test performed already will not be repeated.
The machine data obtained from the document processor may include machine operational data such as set point, actuator, and sensor data collected during regular operation of the machine, machine usage data, historical data such as fault counters and performance data. Such data may be obtained directly from the print engine 110, document feed 120, and other components of the document processor 100, from the sensor array 160, or from the memory 150. In addition, a set of special diagnostic or performance tests may be run on the document processing machine 100 and the effects of the tests can be observed on sensor array 120. Examples of such tests are described in U.S. Pat. Nos. 5,864,730; 5,893,008; 5,903,796; 5,937,224; 5,960,228; 5,946,521; 5,995,775; 6,016,204; 6,081,348; and 6,198,885, commonly owned with this application, the disclosure of which being incorporated herein in its entirety. The diagnostic tests typically are stress tests on the various components that for example, vary the actuators between maximum and minimum values. Examples of such diagnostic tests include but are not limited to photoreceptor uniformity tests, cleaner stress test, charge system stress tests, exposure system tests, development system tests, banding test on the photoreceptor and reload test on the photoreceptor. This data can be supplied to the machine data module 7 and may be factored into the analysis performed by diagnostic engine 2.
Machine data module 7 is added to the diagnostic system for this purpose and provides the mechanism to collect and deliver machine data to diagnostic preprocessor 3 for consideration in the analysis performed by diagnostic engine 2. The machine data processor 19 generates a set of high level features from the machine data based on analysis of the machine data. The generated features diagnostic are converted to a format acceptable to the particular diagnostic engine used, for example event format, in diagnostic preprocessor 3. Some sample machine data features are listed below.
Examples of machine data features and their correlation to malfunctions are shown in
The algorithms for extraction of the machine data features from the raw machine data are based on statistical analysis techniques, including, discriminant analysis, classifiers and regression analysis, and data mining techniques, including, decision trees and the like. The parameters of these tests are predetermined by analysis of machine data corresponding to the various malfunctions as well as data from a normal machine with no malfunctions. These parameters are stored in the machine data processor. The machine data generated during any run of the proposed diagnostic system is analyzed using these predetermined parameters and the feature values are determined.
Processing in preprocessor 3 may also involve conversion of analog data, received from one or more sensors, into qualitative values. Alternatively, it may involve translating machine signals into discrete event sequences, as described in U.S. application Ser. No. 60/154,016, commonly owned with this application, the disclosure of which being incorporated herein in its entirety. The event sequences are recognized by the diagnostic engine 2.
The features generated by the machine data processor and formatted by the diagnostic preprocessor are passed on to the diagnostic engine 2 to be used in conjunction with the defect features for isolating the malfunction.
Machine data module 7 may take on a variety of architectures depending on the means by which the data is collected, an example of such a module is shown in FIG. 4. In
The output of the diagnostic engine 2 is received by the repair planning module 5 for selection of the appropriate service procedure which will correct the malfunction identified in analysis. An example of repair planning module 5 is shown for illustration in FIG. 5. Memory 15 contains a table of repair procedures correlated with specific malfunctions which can be presented for selection by repair planning processor 12. In certain instances, it may be advantageous to review the scheduled maintenance service procedures stored in memory 16 for the purpose of accelerating such service to fix the malfunction. In addition memory 17 provides access to a systems history to allow the replacement of parts where warranted. The repair procedures may be performed by the customer, system administrator, key operator, or the like, or by certified service personnel.
It should be noted that the above description identifies the components of the diagnostic system as separate functional modules. These functions can be implemented by a wide variety of computer elements, for example, a special purpose computer, a programmed microprocessor or microcontroller, a peripheral integrated circuit element, such as an ASIC, a digital signal processor, a hard-wired electronic or logic circuit, such as a discreet element circuit, a programmable logic device such as a PLD, PLA, FPGA, PAL, or in general, any device capable of implementing the functions described herein.
The basic operation of the system of this application involves performing banding defect analysis on a print or copy sample of a special test pattern. Based on the output of the banding defect analysis, the presence of a defect (or set of defects) is detected and further, the defects are characterized by a set of quantitative parameters. Further processing of the banding defect analysis outputs is performed to extract the key features of the banding defects. The extraction of said key features is based on a a series of algorithms which utilize threshold values of the features and device specific specifications.
The key features are analyzed in a diagnostic engine. This results in the computation of the most likely malfunctions among various malfunctions for a given set of key feature values. The process of analysis can be repeated through iterations designed to isolate the most probable malfunction based on a refined set of key feature values. The iterative process can consider image data, machine data, or user input data. A predetermined value of certainty can be set for the purpose of comparison and screening of the probable malfunctions. The analysis is deemed successful upon the selection of a malfunction having a certainty above a preset threshold certainty level. From this data, a suitable service procedure is selected designed to correct the malfunction. If the analysis fails to achieve the threshold of certainty, then a list of most probable causes given the analysis up to that point is displayed with suitable service procedure recommendations. If no malfunction is perceived, then the diagnostic service is aborted and the user is notified.
As shown in
The analysis process can cycle through a series of iterative steps by having a variety of test patterns available in storage. In addition the test patterns can be run in different settings or using different components to isolate the malfunction. Instructions can be presented to the user through the user interface or the cycles can be accomplished automatically as instructed by the diagnostic controller 4.
The diagnostic data can be supplemented by machine data as previously described. As shown in the steps of
In this manner a system and method of servicing a document processing machine is provided without relying on a service technician or other external process. This will result in considerable savings of time and cost. The system and method described herein can readily be adapted for significant automatic control.
While the invention has been described with reference to specific embodiments, the description of the specific embodiments is illustrative only and is not to be construed as limiting the scope of the invention. Various other modifications and changes may occur to those skilled in the art without departing from the spirit and scope of the invention as defined by the claims.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4335952 | Conly et al. | Jun 1982 | A |
| 5138377 | Smith et al. | Aug 1992 | A |
| 5365310 | Jenkins et al. | Nov 1994 | A |
| 5884118 | Mestha et al. | Mar 1999 | A |
| 6275600 | Banker et al. | Aug 2001 | B1 |
| 6504625 | Amero et al. | Jan 2003 | B1 |
| 6522430 | Dalal et al. | Feb 2003 | B1 |
| 6529616 | Rasmussen et al. | Mar 2003 | B1 |
| 6535865 | Skaaning et al. | Mar 2003 | B1 |
| 6571000 | Rasmussen et al. | May 2003 | B1 |
| 6597473 | Rasmussen et al. | Jul 2003 | B1 |
| 6606395 | Rasmussen et al. | Aug 2003 | B1 |
| 6665425 | Sampath et al. | Dec 2003 | B1 |
| Number | Date | Country | |
|---|---|---|---|
| 20030142985 A1 | Jul 2003 | US |
