Abnormal printing fault recognition and laser printer

Abstract

The invention relates to the technical field of laser printers, and discloses an abnormal printing fault recognition method and a laser printer, and the method comprises the steps: obtaining the printing data of the laser printer, carrying out the preprocessing, carrying out the abnormal detection of the data through a COF algorithm, and obtaining the abnormal degree and connection path of each data point. And further analyzing neighborhood data point distribution of each data point, calculating a neighborhood deviation degree, and evaluating a clutter degree and path uniformity in a path direction. And taking the ratio of the neighborhood deviation degree to the path uniformity as an adjustment factor, calculating the significant anomaly degree in combination with the anomaly degree of the data point, performing anomaly detection based on the significant anomaly degree of each data point, and identifying the data anomaly in the working process of the laser printer. According to the method, the COF algorithm, the neighborhood deviation degree and the path uniformity are comprehensively evaluated, so that the data exception in the working process of the laser printer is more accurately identified, and the accuracy and the reliability of exception detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of laser printers, and particularly to an identification of printing abnormal faults and a laser printer. Background Art

[0002] A laser printer is a common office device that transfers image or text information onto a photosensitive drum through laser scanning technology and then completes printing through steps such as developing, transferring, and fixing. Its main advantages include fast printing speed, high quality, and suitability for large-scale printing. However, various faults may occur in laser printers, such as blurred printing, background ash, ghosting, black bands, etc., and these problems are usually related to the states of components such as the photosensitive drum, toner, and charging roller.

[0003] Accurately identifying printing abnormal faults is the key to ensuring the normal operation of the printer. Fault identification can help users quickly locate problems and narrow down the troubleshooting scope, thereby improving the maintenance efficiency. For example, blurred printing may be caused by a damp photosensitive drum or insufficient toner, while background ash in printing may be related to an aging photosensitive drum or too low developing bias voltage. By promptly identifying abnormalities, further damage to the device can be avoided and the service life can be extended.

[0004] Currently, when identifying printing abnormalities in a laser printer through the COF algorithm, there may be problems in the processing of boundary data points. Since most of the neighborhood data points of the boundary data points are on the same side and the distances between these data points are small, the degree of abnormality of the boundary data points is underestimated. This calculation deviation will affect the accuracy of abnormality identification and thus reduce the reliability of fault diagnosis. Summary of the Invention

[0005] The objective of the present invention is:

[0006] To solve the problem that when identifying printing abnormalities in a laser printer through the COF algorithm, most of the neighborhood data points of the boundary data points are on the same side and the distances are small, resulting in the underestimation of their degree of abnormality, which in turn affects the accuracy of abnormality identification and the reliability of fault diagnosis. The present invention provides an identification of printing abnormal faults and a laser printer.

[0007] A method for identifying printing abnormal faults, comprising: obtaining printing data of a laser printer and performing preprocessing; wherein, the printing data includes fixing temperature, preheating time, main motor speed, driving circuit current, toner consumption speed, and print head movement data; performing abnormal detection on the preprocessed printing data according to the COF algorithm to obtain the abnormal degree of each data point and the connection path; obtaining the neighborhood deviation degree according to the distribution of the neighborhood data points of each data point and the Euclidean distance between the actual neighborhood center and the center point, calculating the cosine distance of each segment in the connection path, analyzing the similarity in direction of each segment to obtain the clutter degree of the path direction, and based on the clutter degree, calculating the average distance between the neighborhood data points and the actual center point to evaluate the path uniformity of the neighborhood data points; using the ratio of the neighborhood deviation degree and the path uniformity as an adjustment factor, and multiplying it by the abnormal degree of the data point to obtain the significant abnormal degree of each data point, and performing abnormal detection based on the significant abnormal degree of each data point to identify data anomalies during the operation of the laser printer.

[0008] The effect is that: by obtaining the printing data of the laser printer and performing preprocessing, using the COF algorithm to perform abnormal detection on the preprocessed data to obtain the abnormal degree and connection path of each data point. Further analyzing the distribution of neighborhood data points, neighborhood deviation degree, clutter degree of path direction, and path uniformity, using the ratio of the neighborhood deviation degree to the path uniformity as an adjustment factor, combining it with the abnormal degree of the data point to obtain the significant abnormal degree of each data point. This method can effectively solve the problem that the abnormal degree of boundary data points is underestimated, improve the accuracy of abnormal identification and the reliability of fault diagnosis, thereby reducing maintenance costs and improving the operating efficiency of the laser printer.

[0009] Preferably, obtaining the abnormal degree of each data point includes:

[0010] Calculating the distance between each data point and all other data points to generate a distance matrix;

[0011] Presetting the number of neighborhood data points, taking any data point as the target data point, starting from the target data point, selecting the shortest path among the neighborhood data points to construct a connection path;

[0012] Obtaining the distance value of each step in the connection path, assigning weights to the distances of each step according to the step sequence, and performing weighted summation and dividing by the sum of the weights to obtain the local average link distance, calculating the ratio between the local average link distance of the target data point and the average value of the local average link distances of the neighborhood data points to obtain the abnormal degree of each data point.

[0013] Preferably, obtaining the connection path includes:

[0014] Select a preset number of neighborhood data points in the distance matrix. Starting from any data point, obtain the data point with the shortest distance among the nearest data points to the starting point. Traverse all neighboring data points, obtain the shortest path trajectory from the starting point to each neighborhood point, and construct a connection path.

[0015] Preferably, the neighborhood deviation degree includes:

[0016] Taking any item of data of each data point in the print data as the target item data, calculate the difference between the target item data and the average value of the neighborhood data, and take the ratio of the difference to the standard deviation as the standard deviation degree;

[0017] Calculate the ratio between the number of neighborhood data greater than the target item data and the number of neighborhood data less than the target item data in the neighborhood data of the target item data respectively, and take the absolute difference between 1 and the ratio as the distribution asymmetry;

[0018] Multiply the mean value of the standard deviation degrees of all item data by the mean value of the distribution asymmetries of all item data to obtain the neighborhood deviation degree of the data point.

[0019] The effect is that by calculating the difference between the target item data of each data point and the average value of the neighborhood data, and taking the ratio of the difference to the standard deviation as the standard deviation degree, and at the same time calculating the ratio of the number of neighborhood data greater than and less than the target item data to evaluate the distribution asymmetry, and finally combining the mean value of the standard deviation degree and the mean value of the distribution asymmetry, the neighborhood deviation degree of the data point is obtained. This method can effectively evaluate the degree to which the data point deviates from its neighborhood pattern and improve the accuracy of anomaly detection.

[0020] Preferably, the clutter degree includes:

[0021] Taking any data point in the print data as the target data point, selecting any neighborhood data point in the connection path corresponding to the target data point as the marked neighborhood data point, calculating the vector of the data adjacent to the marked neighborhood data point, calculating the vectors of the other neighborhood data points except the marked neighborhood data point and the adjacent data, and calculating the sum of the products of the vectors of all marked neighborhood data points and other neighborhood data points to obtain the neighborhood direction consistency of the marked neighborhood data point;

[0022] Calculate the Euclidean distances between the marked neighborhood data point and other neighborhood data points and the previous neighborhood data point respectively, and take the product of the Euclidean distances as the neighborhood distance;

[0023] Subtract the ratio between the neighborhood direction consistency and the neighborhood distance from 1 to obtain the clutter degree of the marked neighborhood data point.

[0024] Its effect is as follows: By calculating the sum of the vector products of the marked neighborhood data points and other neighborhood data points, the consistency of the neighborhood direction is evaluated, and combined with the neighborhood distance, the clutter degree of the marked neighborhood data points is calculated, so as to effectively evaluate the complexity and irregularity of the neighborhood path of the data points.

[0025] Preferably, the path uniformity includes:

[0026] Taking any data point in the print data as the target data point and any item of data of the target data point as the target item data, calculating the average value of the neighborhood data of the target item data, and taking the average value of the neighborhood data of all items of data of the target data point as the actual center point;

[0027] Obtaining the sum of the clutter degrees between the vectors corresponding to each neighborhood data point of the target data point and the vectors corresponding to other neighborhood data points, and taking the smallest sum of the clutter degrees as the reference vector of the target data point;

[0028] Taking the cosine distance between the vector between any neighborhood data point of the target item data and the previous data point and the reference vector as the path direction consistency;

[0029] Taking the Euclidean distance between any neighborhood data point of the target item data and the actual center point of the neighborhood data of the target data point as the neighborhood degree;

[0030] According to the average value of the neighborhood degrees and the average value of the path direction consistencies of all neighborhood data points of the target data point, the path uniformity of the target data point is obtained.

[0031] Preferably, the path uniformity satisfies the following relational expression:

[0032]

[0033] In the formula, H i represents the path uniformity of the i-th data point, K represents the number of neighborhood data points of the data point, S i,k represents the Euclidean distance between the actual center point of the neighborhood data points of the i-th data point and the k-th neighborhood data point, R i represents the maximum Euclidean distance between the i-th data point and the K-th neighborhood data point, P i (k) represents the vector between the k-th neighborhood data point of the i-th data point and its previous neighborhood data point and its subsequent neighborhood data point.

[0034] Preferably, the anomaly detection based on the significant anomaly degree of each data point includes:

[0035] A preset abnormal degree threshold is set. When the abnormal degree of data points continuously appears greater than the abnormal degree threshold for a preset number of times during the operation of the laser printer, the laser printer malfunctions, and the staff needs to be notified in time for maintenance.

[0036] The present invention has the following beneficial effects:

[0037] 1. Through the comprehensive evaluation of the COF algorithm, neighborhood deviation degree, and path uniformity, the present invention can more accurately identify data anomalies during the operation of the laser printer, effectively solve the problem of underestimating the abnormal degree of boundary data points, and improve the accuracy and reliability of anomaly detection.

[0038] 2. The present invention performs anomaly detection based on the significant abnormal degree of each data point, can timely detect potential faults of the laser printer, reduce false alarms and missed alarms, thereby enhancing the reliability of fault diagnosis, ensuring the stable operation of the printer, timely and accurately identifying anomalies and notifying the staff for maintenance, which can avoid the further deterioration of faults, reduce downtime and maintenance costs, improve the overall operation efficiency and reliability of the laser printer, and reduce the total cost of ownership. Description of the Drawings

[0039] Figure 1 It is a flowchart of the method from step S1 to step S4 in a method for identifying printing abnormal faults according to an embodiment of the present invention. Detailed Embodiment

[0040] The following combines the drawings and embodiments to further describe in detail the specific embodiments of the present invention. The following embodiments are used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0041] As Figure 1 shown, an energy-saving big data intelligent analysis method includes steps S1 - S4, specifically as follows:

[0042] S1: Obtain the print data of the laser printer and perform preprocessing;

[0043] Among them, the print data includes but is not limited to: fusing temperature, warm-up time, main motor speed, drive circuit current, toner consumption speed, and print head movement data.

[0044] It should be noted that the preset data acquisition frequency is 1Hz. The print data of the laser printer is acquired, and the jth data collected for the ith time is denoted as D j (i), the total number of acquisitions is denoted as n, and the number of data items collected each time is denoted as N.

[0045] Preprocess the print data of the laser printer. First, remove noise through methods such as moving average method, median filtering, or wavelet transform to reduce random fluctuations and outliers in the data. Then, use means such as linear interpolation, polynomial interpolation, or machine learning model prediction to fill in missing values and handle the missing problems in the data acquisition process. Next, standardize the data using methods such as Z-score standardization or Min-Max standardization to make data with different dimensions comparable.

[0046] S2: Perform anomaly detection on the preprocessed print data according to the COF algorithm to obtain the anomaly degree and connection path of each data point.

[0047] Obtain the anomaly degree of each data point, including:

[0048] Calculate the distance between each data point and all other data points to generate a distance matrix;

[0049] Preset the number of neighborhood data points. Taking any data point as the target data point, start from the target data point and select the shortest path among the neighborhood data points to construct a connection path;

[0050] Obtain the distance value of each step in the connection path, assign weights to the distances of each step according to the step order, and perform weighted summation and divide by the sum of the weights to obtain the local average link distance. Calculate the ratio between the local average link distance of the target data point and the average of the local average link distances of the neighborhood data points to obtain the anomaly degree of each data point.

[0051] Obtain the connection path, including:

[0052] Select a preset number of neighborhood data points in the distance matrix. Taking any data point as the starting point, obtain the data point with the shortest distance among the nearest data points of the starting point, traverse all neighboring data points, and obtain the shortest path trajectory from the starting point to each neighborhood point to construct a connection path.

[0053] In this embodiment, the number of neighborhood data points is 32, and the implementer can adjust it according to the specific situation. Exemplarily, assume that the preset number of neighborhood points is 5, and the nearest neighbors of data point F are {B, C, D, E, G}. F is closest to E, E is closest to D, D is closest to C and G, and C is closest to B. Then the SBN-path of F is {F, E, D, C, G, C, B}, and the corresponding distances are {e1, e2, e3, e4, e5}.

[0054] It should be noted that: Since in the process of performing anomaly detection on print data through the COF algorithm, boundary points need to be specially processed, and their anomaly degrees are adjusted upward to reduce the situation where the uneven distribution of data points in boundary data points makes the anomaly degree too small.

[0055] To prevent the boundary points from affecting the COF algorithm, the neighborhood deviation degree is obtained based on the distribution of the neighborhood data points of each data point, as well as the deviation between the actual center point of the neighborhood data points and the center point, so that the more the neighborhood data points are distributed towards one side and the greater the gap between the actual center point and the center point, the greater the neighborhood deviation degree; then, the path uniformity is obtained based on the distance between the neighborhood data points and their actual center point and the randomness of the direction of each connection path. The specific steps are as follows:

[0056] S3: Obtain the neighborhood deviation degree based on the distribution of the neighborhood data points of each data point and the Euclidean distance between the neighborhood actual center and the center point, calculate the cosine distance of each segment in the connection path, analyze the similarity in direction of each segment, obtain the randomness of the path direction, and based on the randomness, calculate the average distance between the neighborhood data points and the actual center point to evaluate the path uniformity of the neighborhood data points.

[0057] The neighborhood deviation degree includes:

[0058] Taking any item of data of each data point in the print data as the target item data, calculate the difference between the target item data and the average value of the neighborhood data, and take the ratio of the difference to the standard deviation as the standard deviation degree;

[0059] Calculate the ratio between the number of neighborhood data of the target item data that is greater than the target item data and the number that is less than the target item data respectively, and take the absolute difference between 1 and the ratio as the distribution asymmetry;

[0060] Multiply the mean of the standard deviation degrees of all item data by the mean of the distribution asymmetries of all item data to obtain the neighborhood deviation degree of the data point.

[0061] Specifically, the neighborhood deviation degree satisfies the following relational expression:

[0062]

[0063] where B i represents the neighborhood deviation degree of the i-th data point, N represents the data item, D j (i) represents the value of the j-th item data of the i-th data point, represents the average value of the j-th neighborhood data of the i-th data point, σ i,j represents the standard deviation of the j-th item data of the neighborhood data points of the i-th data point, and represent the number of the j-th item data in the neighborhood data points of the i-th data point that is greater than and less than the j-th item data of the i-th data point respectively.

[0064] That is to say, represents the gap between the actual center point of the \(i\)-th data point and the \(i\)-th data point. The larger this gap is, the more the neighboring data points of the \(i\)-th data point deviate from the area centered on the \(i\)-th data point, that is, it has a greater neighborhood deviation degree; the part represents the imbalance of each item of data centered on the \(i\)-th data point. The larger this value is, the greater the deviation degree of the neighborhood.

[0065] It should be noted that in addition to the neighborhood deviation degree, the disorderliness of the directions of each segment of the path and the aggregation degree of the neighboring data points with respect to their actual center points should also be considered to measure whether the printed data is normal.

[0066] The disorderliness includes:

[0067] Taking any data point in the printed data as the target data point, selecting any neighboring data point in the connection path corresponding to the target data point as the marked neighboring data point, calculating the vector of the data adjacent to the marked neighboring data point, calculating the vectors of the other neighboring data points except the marked neighboring data point and the adjacent data, and calculating the sum of the products of the vectors of all marked neighboring data points and other neighboring data points to obtain the neighborhood direction consistency of the marked neighboring data point;

[0068] Calculate the Euclidean distances between the marked neighboring data point and other neighboring data points and the previous neighboring data point respectively, and take the product of the Euclidean distances as the neighborhood distance;

[0069] Subtract the ratio between the neighborhood direction consistency and the neighborhood distance from 1 to obtain the disorderliness of the marked neighboring data point.

[0070] Specifically, the disorderliness satisfies the following relational expression:

[0071]

[0072] In the formula, \(P\) i \((α,β)\) represents the disorderliness of the vectors of the \(α\)-th and \(β\)-th neighboring data points of the \(i\)-th data point and the previous neighboring data point. \(K\) represents the number of neighboring data points of the data point. \(E\) j \((i,α)\), \(E\) j \((i,α - 1)\), \(E\) j \((i,β)\) and \(E\) j \((i,β - 1)\) respectively represent the \(j\)-th item of data of the \(α\)-th, \(α - 1\)-th, \(β\)-th, and \(β + 1\)-th neighboring data points of the \(i\)-th data point. \(d(i,α)\) and \(d(i,β)\) respectively represent the Euclidean distances between the \(α\)-th and \(β\)-th neighboring data points of the \(i\)-th data point and the previous neighboring data point.

[0073] By obtaining the neighborhood data point with the largest sum of cosine distances from other neighborhood data points among all neighborhood data points as the reference data point. If the $i$-th data point is a boundary point, this method can select a reference data point on the side not favored by its neighborhood data points (i.e., the side with sparser data points).

[0074] The path uniformity includes:

[0075] Taking any data point in the printing data as the target data point, taking any item of data of the target data point as the target item data, calculating the average value of the neighborhood data of the target item data, and taking the average value of the neighborhood data of all items of data of the target data point as the actual center point;

[0076] Obtaining the sum of the clutter degrees between the vectors corresponding to each neighborhood data point of the target data point and the vectors corresponding to other neighborhood data points, and taking the smallest sum of clutter degrees as the reference vector of the target data point;

[0077] Taking the cosine distance between the vector between any neighborhood data point of the target item data and the previous data point and the reference vector as the path direction consistency;

[0078] Taking the Euclidean distance between any neighborhood data point of the target item data and the actual center point of the neighborhood data of the target data point as the neighborhood degree;

[0079] According to the average value of the neighborhood degrees and the average value of the path direction consistencies of all neighborhood data points of the target data point, obtaining the path uniformity of the target data point.

[0080] Specifically, the path uniformity satisfies the following relational expression:

[0081]

[0082] In the formula, $H$ i represents the path uniformity of the $i$-th data point, $K$ represents the number of neighborhood data points of the data point, $S$ i,k represents the Euclidean distance between the actual center point of the neighborhood data points of the $i$-th data point and the $k$-th neighborhood data point, $R$ i represents the maximum Euclidean distance between the $i$-th data point and the $K$-th neighborhood data point, $P$ i $(k)$ represents the vector between the $k$-th neighborhood data point of the $i$-th data point and its previous neighborhood data point and its subsequent neighborhood data point.

[0083] That is to say, represents the ratio between the average distance between the neighborhood data points and their actual center and the maximum neighborhood distance $R$ of the $i$-th data point i The larger the ratio, the more dispersed the neighborhood data points are, and the higher the possible path uniformity is;

[0084] It represents the cosine distance between the vectors corresponding to all neighborhood data points and the reference vector. The larger the average value of the cosine distance, the larger the cosine distance between other path vectors and the reference vector. The reference vector is the one with the smallest sum of the cosine distances between all path vectors and other path vectors. The larger the cosine distance between it and other path vectors, the more dispersed the direction distribution of its path vector, indicating that the path is more likely to pass through a larger area and surround the i-th data point.

[0085] S4: Take the ratio of the neighborhood deviation degree and the path uniformity as an adjustment factor, and multiply it by the abnormality degree of the data point to obtain the significant abnormality degree of each data point. Based on the significant abnormality degree of each data point, perform anomaly detection to identify data anomalies during the operation of the laser printer.

[0086] It should be noted that the larger the neighborhood deviation degree, the more likely the data point is a boundary point, and its abnormality degree should be adjusted to a larger value. The larger the path uniformity degree, the more its connection path is around the i-th data point, and the smaller the degree to which its abnormality degree is adjusted upward.

[0087] Preset an abnormality degree threshold. In response to the abnormality degree of the data point being greater than the abnormality degree threshold for a continuous preset number of times during the operation of the laser printer, the laser printer has a running fault and it is necessary to promptly notify the staff for maintenance.

[0088] It should be noted that for the preset abnormality degree threshold T, the empirical value is T = 1. When the abnormality degree of the data point is greater than the abnormality degree threshold for 3 consecutive times during the operation of the laser printer, it is considered that the laser printer has a running fault and it is necessary to promptly notify the staff for maintenance.

[0089] The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art of this technology, without departing from the counting principle of the present invention, several improvements and replacements can be made, and these improvements and replacements should also be regarded as the protection scope of the present invention.

Claims

1. An identification method for printing abnormal faults and a laser printer, characterized in that, Including: Obtain the print data of the laser printer and perform preprocessing; wherein, the print data includes fixing temperature, preheating time, main motor speed, driving circuit current, toner consumption speed, and print head movement data; Perform anomaly detection on the preprocessed print data according to the COF algorithm to obtain the anomaly degree of each data point and the connection path; Obtain the neighborhood deviation degree according to the distribution of the neighborhood data points of each data point and the Euclidean distance between the actual center of the neighborhood and the center point, calculate the cosine distance of each segment in the connection path, analyze the similarity in direction of each segment, obtain the clutter degree of the path direction, and based on the clutter degree, calculate the average distance between the neighborhood data points and the actual center point to evaluate the path uniformity of the neighborhood data points; Take the ratio of the neighborhood deviation degree and the path uniformity as an adjustment factor, and multiply it by the anomaly degree of the data point to obtain the significant anomaly degree of each data point, and perform anomaly detection based on the significant anomaly degree of each data point to identify data anomalies during the operation of the laser printer.

2. The identification of printing abnormal faults and the laser printer according to claim 1, characterized in that, Obtain the anomaly degree of each of the above data points, including: Calculate the distance between each data point and all other data points to generate a distance matrix; Preset the number of neighborhood data points, take any data point as the target data point, start from the target data point, select the shortest path in the neighborhood data points, and construct a connection path; Obtain the distance value of each step in the connection path, assign weights to the distances of each step in the order of steps, perform weighted summation and divide by the sum of the weights to obtain the local average link distance, calculate the ratio between the local average link distance of the target data point and the average value of the local average link distances of the neighborhood data points to obtain the anomaly degree of each data point.

3. The identification of printing abnormal faults and the laser printer according to claim 2, characterized in that, Obtain the connection path, including: Select a preset number of neighborhood data points in the distance matrix, take any data point as the starting point, obtain the data point with the shortest distance among the nearest data points of the starting point, traverse all adjacent data points, obtain the shortest path trajectory from the starting point to each neighborhood point, and construct a connection path.

4. The identification of printing abnormal faults and the laser printer according to claim 1, characterized in that, The neighborhood deviation degree includes: Take any item of data of each data point in the print data as the target item data, calculate the difference between the target item data and the average value of the neighborhood data, and take the ratio of the difference to the standard deviation as the standard deviation degree; Calculate the ratio between the number of neighborhood data of the target item data that is greater than the target item data and the number of neighborhood data that is less than the target item data respectively, and take the absolute difference between 1 and the ratio as the distribution asymmetry; Multiply the mean value of the standard deviation degrees of all item data by the mean value of the distribution asymmetries of all item data to obtain the neighborhood deviation degree of the data point.

5. The identification of printing abnormal failures and the laser printer according to claim 1, characterized in that, The clutter degree includes: Take any data point in the print data as the target data point, select any neighborhood data point in the connection path corresponding to the target data point as the marked neighborhood data point, calculate the vector of the data adjacent to the marked neighborhood data point, calculate the vectors of the other neighborhood data points except the marked neighborhood data point and the adjacent data, and calculate the sum of the products of the vectors of all marked neighborhood data points and the other neighborhood data points to obtain the neighborhood direction consistency of the marked neighborhood data point; Calculate the Euclidean distances between the marked neighborhood data points and other neighborhood data points and the previous neighborhood data point respectively, and take the product of the Euclidean distances as the neighborhood distance; Subtract the ratio between the neighborhood direction consistency and the neighborhood distance from 1 to obtain the clutter degree of the marked neighborhood data points.

6. The identification of printing abnormal faults and the laser printer according to claim 1, characterized in that, The path uniformity includes: Taking any data point in the printing data as the target data point, taking any item of data of the target data point as the target item data, calculating the average value of the neighborhood data of the target item data, and taking the average value of the neighborhood data of all items of data of the target data point as the actual center point; Obtain the sum of the clutter degrees between the vectors corresponding to each neighborhood data point of the target data point and the vectors corresponding to other neighborhood data points, and take the smallest sum of clutter degrees as the reference vector of the target data point; Take the cosine distance between the vector between any neighborhood data point of the target item data and the previous data point and the reference vector as the path direction consistency; Denote the Euclidean distance between any neighborhood data point of the target item data and the actual center point of the neighborhood data of the target data point as the neighborhood degree; Obtain the path uniformity of the target data point according to the average value of the neighborhood degrees of all neighborhood data points of the target data point and the average value of the path direction consistency.

7. The identification of printing abnormal faults and the laser printer according to claim 6, characterized in that, The path uniformity satisfies the following relational expression: Where, H i represents the path uniformity of the i-th data point, K represents the number of neighboring data points of the data point, and S i,k represents the Euclidean distance between the actual center point of the neighboring data points of the i-th data point and the k-th neighboring data point, and R i represents the maximum Euclidean distance between the i-th data point and the K-th neighboring data point, and P i (k) represents the vector between the k-th neighboring data point of the i-th data point and its previous neighboring data point and its next neighboring data point.

8. The recognition of printing abnormal faults and the laser printer according to claim 1, characterized in that, The anomaly detection based on the significant anomaly degree of each data point includes: Preset an anomaly degree threshold. In response to the anomaly degree of the data points exceeding the anomaly degree threshold continuously for a preset number of times during the operation of the laser printer, the laser printer has a running fault and the staff needs to be notified in time for maintenance.

Images