Infrared image enhancement method for machine room monitoring

Abstract

The invention relates to the field of image processing, in particular to an infrared image enhancement method for machine room monitoring, which comprises the following steps: acquiring a preprocessed machine room infrared monitoring image, and uniformly dividing the infrared monitoring image into a plurality of blocks; constructing a corresponding local window by taking each pixel point as a center in the block, and calculating a heat source anomaly index of the pixel point; and according to the heat source anomaly index, calculating to obtain a gain coefficient of the corresponding pixel point during enhancement, taking the machine room infrared monitoring image as input, and performing enhancement processing on the machine room infrared monitoring image based on the gain coefficient. According to the method, the heat source anomaly index of the pixel point is calculated, the gain coefficient is determined according to the heat source anomaly index, and the machine room infrared monitoring image is enhanced, so that key details in the image can be effectively highlighted, parts which are originally blurred or are difficult to distinguish are clearer, and monitoring personnel can more accurately observe the running state of equipment in the machine room.

Description

Technical Field

[0001] The present invention relates to the field of image processing. More specifically, the present invention relates to an infrared image enhancement method for computer room monitoring. Background Art

[0002] The computer room monitoring system is a comprehensive solution to ensure the safety of the computer room and the stable operation of equipment, and the infrared monitoring video system is crucial. This system uses passive infrared thermal imaging or active infrared supplementary lighting technology to achieve 24-hour uninterrupted monitoring by sensing the infrared radiation or reflected light source of objects. Even in extreme environments without visible light, it can generate clear thermal images or near-infrared images, breaking through the lighting limitations of traditional monitoring. This enables the computer room to maintain complete monitoring coverage under complex conditions such as at night, equipment occlusion, or power outage, providing reliable protection for the safety of the computer room.

[0003] The prior art, such as the patent application document with the publication number CN118154491A, discloses a method and device for processing infrared images, a storage medium. The method for processing infrared images first performs compression processing on the original infrared image, and then uses a target detection and instance segmentation model to process the target infrared image, identify and segment the area to be processed containing the target power device, and then perform detail enhancement processing on the area to be processed, including performing gray-scale enhancement and noise suppression operations on the base layer and the detail layer respectively to enhance the detail information of the image.

[0004] However, in a computer room, when the equipment is operating at a low load and the ambient temperature is relatively stable, the difference in infrared radiation intensity between the surface of the equipment and the surrounding environment is small. At this time, for the infrared image processed according to the prior art, its contrast may still not be sufficient for the monitoring personnel to quickly and accurately discover potential abnormal situations, such as hidden faults like local overheating of the equipment, thus affecting the effectiveness of the computer room monitoring system and unable to fully play its role in ensuring the safety of the computer room and the stable operation of equipment. Summary of the Invention

[0005] To solve the above technical problem that it is difficult to effectively highlight the equipment features due to the limitations of infrared image enhancement processing, the present invention provides the following technical solutions.

[0006] An infrared image enhancement method for computer room monitoring, comprising: Obtain the preprocessed infrared monitoring image of the computer room, and evenly divide the infrared monitoring image into multiple blocks; Based on the local windows constructed with each pixel point as the center in the blocks, calculate the heat source anomaly index of the pixel points; Calculate the gain coefficient for each corresponding pixel point during enhancement according to the heat source anomaly index, take the infrared monitoring image of the computer room as the input, and perform enhancement processing on it based on the gain coefficient; Among them, the temperature feature vector of the pixel is calculated, clustering is performed according to the temperature feature vector, pixel connected component analysis is performed on the clustering result to obtain multiple equal feature domains, the area change and temperature change of the equal feature domains to which the same pixel belongs among the selected set number of image frames are compared to obtain the heat source anomaly index of the pixel; the temperature feature vector includes the temperature smoothing index within the local window of the pixel.

[0007] The present invention realizes double verification from local temperature distribution to spatial topological structure by calculating features such as the temperature smoothing index within the local window of the pixel, combined with clustering analysis and connected component analysis; by comparing the area change and temperature change of the equal feature domains to which the same pixel belongs among the selected set number of image frames, it can be judged whether the heat source is abnormal. If the area of the equal feature domain to which a pixel belongs increases sharply within a short time and the temperature also rises significantly, then the heat source anomaly index of this pixel will be very high, which may indicate abnormal situations such as overheating of the computer room equipment; based on the heat source anomaly index, the gain coefficient is calculated to perform differential enhancement on the abnormal area (such as increasing the brightness of the high-temperature area), while retaining the details of the normal area and avoiding image distortion caused by global over-enhancement, so that the monitoring personnel can more intuitively find the location of the heat source and take measures in time.

[0008] Preferably, the temperature feature vector further includes the temperature distribution anomaly index within the local window of the pixel.

[0009] Preferably, the obtaining process of the temperature smoothing index includes: Obtain the gradient magnitude and gradient angle of all pixels of the block; Obtain the number of occurrences of the mode of the gradient angles within the local window of the pixel, and the difference between the gradient angle of each pixel and the mode of the gradient angles, and then calculate the temperature smoothing index within the local window of the pixel according to the gradient magnitude and gradient angle.

[0010] The gradient magnitude can reflect the severity of the temperature change, and the gradient angle represents the direction of the temperature change. Considering the information of the gradient magnitude and gradient angle comprehensively can reflect the smoothness of the temperature distribution in the local area. In the normal heat dissipation area of a device, the temperature change may be relatively gentle and the gradient magnitude is small; while in an abnormal heat source area, the temperature change may be relatively severe and the gradient magnitude is large.

[0011] Preferably, the obtaining process of the temperature distribution anomaly index includes: Divide the gray values and gradients of all pixels included in the local window of the pixel into a set number of gray intervals and gradient intervals respectively; Count the number of pixels in each combination of gray intervals and gradient intervals; Calculate the heat source weight of each pixel in the local window, multiply the total number of pixels in any combination of a gray level interval and a gradient interval by the heat source weights of all pixels in this combination and sum them to obtain the temperature distribution anomaly index.

[0012] By dividing the gray level value (reflecting temperature) and the gradient value (reflecting the temperature change rate) into intervals, the characteristics of both high-temperature regions and regions with abrupt temperature changes can be captured simultaneously. Moreover, by assigning heat source weights to each combination of gray level intervals and gradient intervals, the contribution of abnormal signals can be highlighted and normal background noise can be suppressed.

[0013] Preferably, the clustering is the K-means clustering algorithm.

[0014] Preferably, the process of obtaining the heat source anomaly index includes: Calculate the relative change rate of the pixel points based on the area change and temperature change of the equal feature domains to which the pixel points in the current frame and the previous frame belong; Traverse all the infrared monitoring images of the computer room before the current frame, calculate the relative change rate of the pixel points between adjacent two frames, sum them up and then average to obtain the heat source anomaly index of the pixel points.

[0015] When abnormal heating just starts to occur in the computer room equipment, the temperature change may be relatively small. By calculating the relative change rate between the current frame and the previous frame (combining area change and temperature change), sudden heat source anomalies within a short period can be quickly identified; then traverse all historical frames, sum them up and average to reflect the cumulative effect of heat source anomalies, thereby capturing small temperature changes.

[0016] Preferably, the gain coefficient satisfies the relational expression: ; where represents the gain coefficient when enhancing the pixel points in the current frame , represents the heat source anomaly index of the pixel points in the current frame , represents the average value of the heat source anomaly indexes of all pixels in the current frame.

[0017] Preferably, the relative change rate satisfies the relational expression: ; where represents the relative change rate of the pixel points in the infrared monitoring images of the current frame and the previous frame, is the area of the equal feature domain to which the pixel points belong in the current frame, is the area of the equal feature domain to which the pixel points belonged in the previous frame of the current frame, is the pixel points The gray - scale mean of all pixels within the same feature domain to which it belongs in the current frame, is the pixel point The gray - scale mean of all pixels within the same feature domain to which it belongs in the previous frame of the current frame; where, is a constant, and 1 is used as a hyper - parameter.

[0018] Through the combined change of the spatial range and thermal radiation intensity of the area to which the pixel point belongs, and by combining with the Sigmoid function to achieve non - linear mapping, the spatio - temporal dynamic features of the pixel point are transformed into comparable indicators, which not only improves the positioning accuracy of single - frame anomalies, but also realizes the prediction of the abnormal development trend through temporal accumulation.

[0019] Preferably, the process of obtaining the temperature smoothing index includes: Calculating the entropy of the gray - scale distribution within the local window, and taking the reciprocal of the entropy of the gray - scale distribution within the local window as the temperature smoothing index.

[0020] Preferably, the process of obtaining the temperature distribution anomaly index includes: Calculating the gray - scale mean and gray - scale standard deviation within the local window of the pixel point; Calculating the difference between the gray - scale value of the pixel point and the gray - scale mean within the local window and dividing it by the gray - scale standard deviation to obtain the temperature distribution anomaly index.

[0021] The beneficial effects of the present invention are: The present invention comprehensively considers various factors such as the temperature smoothing index and temperature distribution anomaly index within the local window of the pixel point, and can more comprehensively and accurately evaluate the heat - source anomaly situation of the pixel point. By comparing the area change and temperature change of the same feature domain to which the same pixel point belongs among a set number of selected image frames, the abnormal heating area of the equipment in the computer room can be effectively identified.

[0022] Compared with the traditional image enhancement method, through a series of steps such as feature analysis and clustering, it can more accurately distinguish normal heat sources and abnormal heat sources, reduce the misjudgment rate caused by poor image quality or interference factors, and improve the reliability and accuracy of the monitoring system. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flowchart of the method of steps S1 - S3 in an infrared image enhancement method for computer room monitoring according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments.

[0025] Refer to Figure 1, An infrared image enhancement method for computer room monitoring includes steps S1 - S3, specifically as follows: S1: Obtain the pre - processed infrared monitoring image of the computer room, and evenly divide the infrared monitoring image into multiple blocks.

[0026] In the computer room, when equipment is running, it generates heat, which can be captured by an infrared camera. The infrared camera can detect the thermal radiation of an object and convert it into an image signal.

[0027] For the infrared monitoring video collected by the infrared camera, extract a single - frame infrared monitoring image from the video. In the embodiment of the present invention, the single - frame infrared monitoring image is an 8 - bit grayscale image, which means that the grayscale value range of each pixel point is from 0 to 255. The higher the grayscale value, the higher the thermal radiation intensity of the point, and usually also means a higher temperature.

[0028] Furthermore, perform denoising on the single - frame infrared monitoring image, such as median filtering denoising, to obtain the pre - processed infrared monitoring image of the computer room.

[0029] After the denoising process, perform block processing on the image so that each small block can be analyzed and processed independently. Exemplarily, if the image resolution is 640×480, it can be divided into multiple 80×60 sub - blocks.

[0030] S2: Based on the blocks, construct corresponding local windows centered on each pixel point, and calculate the heat source anomaly index of the pixel point.

[0031] When the equipment in the computer room is working normally, the distribution of its heat source is relatively uniform and stable. However, when the equipment fails or overheats, local bright areas will be formed in the infrared image. The temperature gradient changes greatly in these areas, which is significantly different from the normal heat source distribution pattern. By calculating the heat source anomaly index, it is possible to quantify whether there is an anomaly in the heat source at the position represented by each pixel point, thereby helping to identify these abnormal heat sources.

[0032] Specifically, the thermal radiation of normal equipment has a radially diffusive characteristic, while abnormal heat sources (such as short - circuit, overload) show randomness in direction and drastic changes, that is, the gradient direction is chaotic (the angles are dispersed) and the gradient magnitude changes suddenly (the temperature changes drastically).

[0033] In one embodiment, by quantifying the consistency of the gradient direction and the smoothness of the gradient magnitude within the local window constructed centered on the pixel point, it is determined whether the local window conforms to the thermal radiation pattern of the normal heat source.

[0034] Taking any one of the blocks obtained by dividing the above S1 from the current frame infrared monitoring image as an example, it is used as the target block, and a local window is constructed with any pixel point in the target block as the center. Exemplarily, the size of the local window is set to 7×7 or 9×9, and then the Sobel operator is used to process this block to obtain the gradient magnitude and gradient angle of all pixel points in this block. The range of the gradient angle is from 0° to 360°.

[0035] Further, obtain the number of occurrences of the mode of the gradient angles within the local window of the pixel point, and the difference between the gradient angle of each pixel point and the mode of the gradient angles, and then calculate the temperature smoothing index within the local window of the pixel point according to the gradient magnitude and gradient angle.

[0036] Exemplarily, the above temperature smoothing index satisfies the relational expression:

[0037] is the temperature smoothing index within the local window of pixel point ; is the number of occurrences of the mode of the gradient angles within the local window of pixel point (i.e., the number of main direction pixels), is the total number of pixel points included within the local window of pixel point ; is the absolute value of the difference between the gradient angle of pixel point within the local window of pixel point and the mode of the gradient angles within this local window, is the gradient magnitude of pixel point within the local window of pixel point . Among them, the 1 in the denominator is a hyperparameter used to avoid the denominator being zero.

[0038] Among them, reflects the consistency of the gradient directions within the local window. When the proportion of main direction pixels is high (such as in the edge or smooth area), is close to 1; within the local window, the smaller the difference between the gradient angles of different pixel points and the mode angle, that is, is smaller, indicating that the thermal radiation direction of the heat source is more unified and more in line with the characteristics of a normal heat source. At the same time, the smaller the gradient magnitude of the pixel points within the local window, that is, is smaller, indicating that the temperature gradient change is smoother and more in line with the characteristics of a normal heat source. Therefore, the larger the calculated temperature smoothing index, the smoother the temperature change of this pixel point and its surrounding area, and the more likely it is a normal heat source.

[0039] In one embodiment, the specific characteristics of the temperature distribution are also considered. The heat of a normal heat source gradually diffuses into the surrounding environment along the heat dissipation path. Therefore, in the infrared image, it is manifested as a gradually darkening temperature gradient, that is, the temperature remains within the normal range and the temperature gradient is relatively small. While the heat diffusion range of an abnormal heat source may be small and the heat is concentrated in a local area. Therefore, in the infrared image, it is manifested as an isolated high-brightness area, that is, the temperature exceeds the normal range and the temperature gradient is relatively large.

[0040] Still taking any one of the blocks obtained by the above S1 division as an example, it is used as the target block, and a local window of the same size as above is constructed with any pixel point in the target block as the center.

[0041] Furthermore, the gray value range of all pixel points included in the pixel point local window is evenly divided into, for example, A = 5 gray intervals, and at the same time, the gradient magnitude range of all pixel points included in the pixel point local window is evenly divided into B = 6 gradient intervals.

[0042] The number of pixels in each combination of gray intervals and gradient intervals is counted to form a matrix of A×B. For example, represents the number of pixels with gray level in the a-th interval and gradient in the b-th interval.

[0043] Next, calculate the heat source weight of each pixel point in the local window, and then multiply the above by the heat source weight of the corresponding pixel point and sum to obtain the temperature distribution anomaly index in this local window.

[0044] Then the above temperature distribution anomaly index satisfies the relational expression:

[0045] In the formula, is the temperature distribution anomaly index in the local window of pixel point , represents the interval index of the gray value in this local window, and the value range is from 1 to A; represents the number of intervals of the pixel gray value in this local window; represents the interval index of the gradient in this local window, and the value range is from 1 to B; represents the number of intervals of the pixel gradient magnitude in this local window; is the heat source weight of the pixel with gray level in the a-th interval and gradient in the b-th interval, reflecting the comprehensive influence of gray level and gradient.

[0046] Among them, increases with the increase of the gray interval and the gradient interval, strengthens the contribution of the "high gray level + high gradient" area, and makes the anomaly index more sensitive to areas with sharp changes or mutations at the edge.

[0047] If the pixel points with larger gray values and gradient values are distributed in the lower right corner of the matrix, that is, when a and b are larger, the corresponding is larger, indicating that the temperature in this window is higher and the temperature change is larger, and the abnormal degree of the heat source distribution is greater. Therefore, the calculated temperature distribution anomaly index is larger.

[0048] In summary, the temperature smoothing index and the temperature distribution anomaly index evaluate the temperature distribution in the local window from different perspectives. The temperature smoothing index mainly focuses on the smoothness of temperature changes, while the temperature distribution anomaly index pays more attention to the specific characteristics of the temperature distribution. By combining these two indexes, the temperature distribution in the local window can be evaluated more comprehensively, so as to more accurately identify and distinguish normal heat sources and abnormal heat sources.

[0049] According to the above operations, the temperature smoothing index and the temperature distribution anomaly index in the local windows of all pixel points in the target block can be calculated in the same way. Further, the temperature smoothing index and the temperature distribution anomaly index in the local windows of all pixel points on the infrared monitoring image can be obtained in the same way.

[0050] In another embodiment, the entropy of the gray distribution in the local window can also be calculated, and the reciprocal of the entropy of the gray distribution in the local window is used as the temperature smoothing index.

[0051] In another embodiment, a method for simply calculating the above temperature distribution anomaly index is provided, that is: Calculate the gray mean value and gray standard deviation in the local window of the pixel point; calculate the difference between the gray value of the pixel point and the gray mean value in the local window and divide it by the gray standard deviation to obtain the temperature distribution anomaly index.

[0052] In one embodiment, the above temperature smoothing index and temperature distribution anomaly index are combined into a temperature feature vector to describe the temperature change pattern of the pixel point. Taking the temperature feature vectors of all pixel points as input, the K-means clustering algorithm is used for clustering. Exemplarily, K is taken as 4. After clustering, the pixel points belonging to the same clustering cluster have similar temperature change characteristics.

[0053] Perform connected component analysis on the pixel points belonging to the same clustering cluster to obtain several connected components, and these connected components are called equal feature domains. The pixel points within the equal feature domain have similar temperature change characteristics.

[0054] It should be noted that in the infrared monitoring video of the computer room, the monitoring position and the positions of all equipment in the computer room are fixed. Therefore, in two adjacent frames of infrared images of the computer room, the actual positions represented by the pixel points at the same position are the same or extremely close.

[0055] If in the infrared images of the computer room for two adjacent frames, the temperature of the pixels at the same position in the same characteristic domain rises rapidly and the heating area becomes larger, it indicates that there may be an abnormality in the heat source at this time, such as local overheating caused by a fault or local heat release due to a short circuit in the circuit. Therefore, for each pixel in the current frame of the infrared monitoring image, compare the area change and temperature change of the characteristic domain to which it belongs in the previous M frames (such as the previous 20 frames), so as to calculate the heat source abnormality index of the pixel.

[0056] Exemplarily, first calculate the relative change rate between the current frame and the previous frame, that is, the relational expression is satisfied as:

[0057] In the formula, represents the relative change rate of the pixel in the infrared monitoring images of the current frame and the previous frame, is the area of the characteristic domain to which the pixel belongs in the current frame, is the area of the characteristic domain to which the pixel belongs in the previous frame of the current frame, is the average gray value of all pixels in the characteristic domain to which the pixel belongs in the current frame, is the average gray value of all pixels in the characteristic domain to which the pixel belongs in the previous frame of the current frame. Among them, is a constant (usually taking a minimum value to ensure that the denominator is not zero), and 1 is a hyperparameter.

[0058] Among them, combines the influence area (spatial dimension) and temperature level (intensity dimension) of the heat source, and is used to measure the overall abnormality degree of the heat source. For example, if the temperature of a certain area rises slightly but the area expands significantly (such as poor heat dissipation), or the area remains unchanged but the temperature rises suddenly (such as short-circuit heat release), both may cause to increase. The larger the ratio, the more significant the abnormality. Further, after subtracting 1 from the ratio, it is mapped to the [0, 1] interval through the Sigmoid function, and a large gradient is generated when the ratio is close to 1 (that is, a small change), so as to amplify the weak abnormality signal and improve the detection sensitivity.

[0059] Finally, traverse the previous M frames of infrared monitoring images of the current frame, calculate the relative change rate between adjacent two frames of pixels and sum and average them to obtain the heat source abnormality index of the pixel

[0060] ​Generally speaking, the above calculation of the heat source anomaly index for pixels depends on the clustering results (such as the division of feature domains) of the temperature smoothing index and the temperature distribution anomaly index, so as to realize the mapping from spatial features to the time dimension. The larger the temperature smoothing index, the smoother the temperature change in the area, and the more it conforms to the radial diffusion characteristics of normal heat sources. Since normal heat sources (i.e., high temperature smoothing indices) usually do not have significant abnormal changes, the heat source anomaly index is relatively low; the temperature distribution anomaly index reflects the persistence of heat source anomalies. A high temperature distribution anomaly index (presence of abnormal hotspots) may lead to an increase in the heat source anomaly index because the abnormal hotspots persist or expand over time.

[0061] S3: Calculate the gain coefficient for the corresponding pixel during enhancement according to the heat source anomaly index. Take the infrared monitoring image of the computer room as the input and perform enhancement processing on it based on the gain coefficient.

[0062] In the infrared image of the computer room, pixels of abnormal heat sources (such as equipment overload, line short - circuit, etc.) need to be enhanced preferentially so that the operation and maintenance personnel can quickly locate the problem; reduce the enhancement of areas with normal temperature to avoid excessive interference information.

[0063] By comparing the heat source anomaly indices, calculate independent gain coefficients for each pixel to achieve the effect of "more enhancement in abnormal areas and less enhancement in normal areas".

[0064] Then the gain coefficient of the pixel calculated according to the heat source anomaly index satisfies the relational expression:

[0065] In the formula, represents the gain coefficient of the current - frame pixel during enhancement, represents the heat source anomaly index of the current - frame pixel and represents the average value of the heat source anomaly indices of all pixels in the current frame.

[0066] Furthermore, take each frame of the infrared monitoring image of the computer room as the input. For each pixel in the image, calculate its gain coefficient according to the above formula , and use the calculated gain coefficient to perform enhancement processing on each pixel to highlight the abnormal heat source part. Finally, generate an enhanced infrared image of the computer room, which is convenient for the monitoring personnel to more clearly identify the abnormal heat source and timely investigate potential safety hazards.

[0067] If the heat source anomaly index of a certain pixel is greater than the average of the heat source anomaly indexes of all pixels, it indicates that the degree of heat source anomaly at the position represented by this pixel is higher than the overall heat source anomaly in the image. In this case, this point is more likely to be an abnormal heat source and needs to be enhanced more significantly; conversely, if the heat source anomaly index of a certain pixel is less than the average of the heat source anomaly indexes of all pixels, it indicates that the degree of heat source anomaly at the position represented by this pixel is lower than the overall heat source anomaly in the image. In this case, this point is more likely to be a normal heat source point and requires less enhancement.

[0068] Thus, the enhancement process of the infrared image in the computer room monitoring is completed.

[0069] It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent shall be subject to the appended claims.

Claims

1. An infrared image enhancement method for computer room monitoring, characterized in that Including: Obtain the pre - processed infrared monitoring images of the computer room, and evenly divide the infrared monitoring images into multiple blocks; Based on the corresponding local windows constructed with each pixel point as the center in the block, calculate the heat source anomaly index of the pixel point; Calculate the gain coefficient for corresponding pixel points during enhancement according to the heat source anomaly index, take the infrared monitoring image of the computer room as the input, and perform enhancement processing on it based on the gain coefficient; Among them, calculate the temperature feature vector of the pixel point, perform clustering according to the temperature feature vector, perform pixel point connected - component analysis on the clustering result to obtain multiple equal - feature domains, compare the area change and temperature change of the equal - feature domains to which the same pixel point belongs among the set number of selected image frames, and obtain the heat source anomaly index of the pixel point; the temperature feature vector includes the temperature smoothing index within the local window of the pixel point.

2. The infrared image enhancement method for computer room monitoring according to claim 1, wherein, The temperature feature vector also includes the temperature distribution anomaly index within the local window of the pixel point.

3. An infrared image enhancement method for computer room monitoring according to claim 2, characterized in that The process of obtaining the temperature smoothing index includes: Obtain the gradient magnitude and gradient angle of all pixel points in the block; Obtain the number of occurrences of the mode of the gradient angles within the local window of the pixel point, and the difference between the gradient angle of each pixel point and the mode of the gradient angles, and then calculate the temperature smoothing index within the local window of the pixel point according to the gradient magnitude and gradient angle.

4. An infrared image enhancement method for computer room monitoring according to claim 3, characterized in that, The process of obtaining the temperature distribution anomaly index includes: Evenly divide the gray values and gradients of all pixel points included in the local window of the pixel point into a set number of gray intervals and gradient intervals respectively; Count the number of pixels in each combination of gray intervals and gradient intervals; Calculate the heat source weight of each pixel point within the local window, multiply the total number of pixels in any combination of gray intervals and gradient intervals by the heat source weights of all pixel points within this combination and sum to obtain the temperature distribution anomaly index.

5. An infrared image enhancement method for computer room monitoring according to claim 4, characterized in that, The clustering is the K - means clustering algorithm.

6. An infrared image enhancement method for computer room monitoring according to claim 5, characterized in that, The process of obtaining the heat source anomaly index includes: Calculate the relative change rate of the pixel point based on the area change and temperature change of the equal - feature domains to which the pixel point belongs in the current frame and the previous frame; Traverse all infrared monitoring images of the computer room before the current frame, calculate the relative change rate between adjacent two frames of pixel points and sum and average them to obtain the heat source anomaly index of the pixel point.

7. An infrared image enhancement method for computer room monitoring according to claim 6, characterized in that The gain coefficient satisfies the relation: ; In the formula, represents the pixel points of the current frame is the gain coefficient during enhancement, represents the heat source anomaly index of the pixel points of the current frame and represents the average value of the heat source anomaly indices of all pixels in the current frame.

8. An infrared image enhancement method for computer room monitoring according to claim 7, characterized in that, The relative change rate satisfies the relation: ; In the formula, represents the pixel point the relative change rate in the current frame and the previous frame of the infrared monitoring image, is the area of the equal feature domain to which the pixel point belongs in the current frame, is the area of the equal feature domain to which the pixel point belongs in the previous frame of the current frame, is the average gray value of all pixels within the equal feature domain to which the pixel point belongs in the current frame, is the average gray value of all pixels within the equal feature domain to which the pixel point belongs in the previous frame of the current frame; where is a constant, and 1 is used as a hyperparameter.

9. An infrared image enhancement method for computer room monitoring according to claim 3, characterized in that The process of obtaining the temperature smoothing index includes: Calculate the entropy of the gray - level distribution within the local window, and take the reciprocal of the entropy of the gray - level distribution within the local window as the temperature smoothing index.

10. An infrared image enhancement method for computer room monitoring according to claim 3, characterized in that, The process of obtaining the temperature distribution anomaly index includes: Calculate the gray - level mean and gray - level standard deviation within the local window of the pixel point; Calculate the difference between the gray - level value of the pixel point and the gray - level mean within the local window and divide it by the gray - level standard deviation to obtain the temperature distribution anomaly index.

Images