A video monitoring system and method for a substation

Abstract

The present invention provides a video monitoring system and method for a substation. The system includes an image acquisition device, a monitoring main end and a monitoring mobile end, and the image acquisition device and the monitoring mobile end are both communicatively connected to the monitoring main end; the method includes: obtaining video image data of the monitored area in real time, selecting a preset number of frames in the preprocessed image as selected frames, and generating a video sequence with the selected frames in units of frames; performing de-jitter processing on the video sequence through the inter-frame correlation coefficient; performing moving object detection on the video sequence after de-jitter processing, and when a moving object is detected in the video, prompting through an information display module that there is a moving object in the corresponding video, and at the same time sending the information of the appearance of the moving object to the monitoring mobile end; through the system and method, it is possible to maintain a good recognition rate of moving objects even when the background environment of the video image changes.

Description

Technical Field

[0001] The present invention relates to the technical field of video surveillance, and in particular, to a video surveillance system and method for a substation. Background Art

[0002] A substation is a place in the power system where the voltage and current of electric energy are transformed, concentrated, and distributed. To ensure the quality of electric energy and the safety of equipment, voltage adjustment, tidal control, and protection of transmission and distribution lines and major power equipment are usually also carried out in the substation. It can be seen that the substation is crucial in the power system.

[0003] China has a vast territory. With the continuous improvement of the power system infrastructure, the number of substations is also extremely large. Due to the importance of substations in the power system, the management of substations is particularly important. However, as the number of substations increases, the invested human resources also become more and more large, which causes a relatively large cost burden on power enterprises. With the update of technology, intelligent technology has been continuously integrated into the management work of substations. In recent years, unmanned intelligent substations have emerged more and more. Power enterprises have begun to achieve unmanned operation of substations through remote video surveillance, and only send staff to deal with them when a fault or abnormality occurs, which effectively controls the labor cost for power enterprises.

[0004] Existing unmanned substations mainly achieve inspection through video surveillance. Even some large substations are equipped with inspection robots. Among them, the detection of moving objects is mainly used in video surveillance. Currently, the commonly used methods for detecting moving objects include frame difference method, background subtraction method, etc. Among them, the frame difference method is a method with a relatively simple implementation, requiring less computational effort. It only needs to intercept three consecutive frames of images in the video sequence and perform subtraction operations on these three frames of images to achieve. However, its disadvantages are also very obvious. Since the method is relatively simple, it does not consider environmental changes such as illumination in the video sequence. The background subtraction method is a currently commonly used method for detecting moving objects. The background subtraction method needs to first construct an accurate background image, and then discover the foreground through the difference between the test image and the background image. Therefore, in the background method, how to obtain an accurate background image is very important. And if the background image cannot be updated in time, its detection effect will also be greatly reduced over time.

[0005] In summary, we believe that in current intelligent substations, there is still room for further improvement in the method for detecting moving objects based on video surveillance images. Summary of the Invention

[0006] The object of the present invention is to provide a video monitoring system and method for a substation, which can still have good moving object detection ability in a changing background environment such as light and shadow, sand and dust weather, and help to implement the inspection work of the substation.

[0007] The embodiments of the present invention are realized through the following technical solutions:

[0008] In a first aspect, a video monitoring system for a substation is provided, including an image acquisition device, a monitoring main end and a monitoring mobile end. The image acquisition device and the monitoring mobile end are both communicatively connected to the monitoring main end. The image acquisition device is used to collect environmental and instrument image data and send it to the monitoring main end;

[0009] The monitoring main end includes a processing and control module and an information display module. The processing and control module includes an image processing unit and an abnormal warning unit. The image processing unit is used to obtain and process the data collected by the image acquisition device. Specifically, it obtains the first frame image in the video sequence, calculates the background dynamic coefficient and initializes the background model with it, obtains the next frame image as the detected frame, performs foreground detection on the detected frame according to the background model, then updates the background model through the detected frame and uses the updated background model to detect the next detected frame in the video sequence. Repeat the above process until all frames in the current video sequence are detected. If there is a foreground in the detected frame, it is determined that a moving object appears. The abnormal warning unit is used to give a warning according to the processing result of the image processing unit. Specifically, when the image processing unit detects a moving object, the abnormal warning unit gives a warning prompt. The information display module is used to display the data collected by the image acquisition device and highlight the data of the detected moving object;

[0010] The monitoring mobile end is used to view the images collected by the image acquisition device through the network and prompt the moving object detection information.

[0011] Furthermore, the information display module includes a resource management and control unit, a configuration unit, a video visualization unit, a service unit and a network unit. The resource docking unit is used to obtain the data output by the image acquisition device and the image processing unit and display it in the resource tree. The configuration unit is used to complete the local configuration of the information display module itself. The video visualization unit is used to complete the local video preview and playback operations of the information display module. The service unit includes a service configuration unit and a service control unit. The service configuration unit is used to complete the configuration and decoding channel association of the information display module. The service control unit is used to complete the window opening, splitting, roaming and wall mounting processing of the display window. The network unit is used to complete the request and response processing of messages.

[0012] Further, the processing control module further includes a jitter processing unit, which is configured to use the middle part of the initial frame image of the video sequence after removing a preset number of pixels from the top, bottom, left, and right as a reference frame, perform horizontal and vertical projections on the reference frame and the relevant frames, then describe the inter-frame correlation through the Pearson correlation coefficient, select the region with the largest correlation between the relevant frame and the reference frame, thereby realizing the repair of jitter, and obtaining the repaired video sequence.

[0013] In a second aspect, a video monitoring method for a substation is provided. The video monitoring method includes:

[0014] Real-time acquiring video image data of the monitored area, selecting a preset number of frames in the preprocessed image as selected frames, and generating a video sequence with the selected frames in units of frames;

[0015] Performing de-jitter processing on the video sequence through the inter-frame correlation coefficient;

[0016] Performing moving object detection on the de-jitter processed video sequence. When a moving object is detected in the video, the information display module is used to prompt that there is a moving object in the corresponding video, and at the same time, the information of the appearance of the moving object is sent to the monitoring mobile terminal.

[0017] Further, the step of selecting a preset number of frames in the preprocessed image as selected frames and generating a video sequence with the selected frames in units of frames is specifically as follows: uniformly acquiring a preset number of frames in the preprocessed image as selected frames at an average time interval or an average frame interval, and the preprocessed image is extracted from the video data at a preset time interval or a preset frame interval.

[0018] Further, the step of performing de-jitter processing on the video sequence through the inter-frame correlation coefficient is specifically as follows: preprocessing the video sequence, and using the image obtained by removing a preset number of pixels from the top, bottom, left, and right of the first frame image in the video sequence as a reference frame;

[0019] Performing grayscale processing on all frames in the video sequence, and then performing local adaptive binarization processing on the grayscale processed image;

[0020] Performing horizontal and vertical projection calculations on the binarized image, then calculating the quotient of the covariance and the standard deviation of each frame and the reference frame in different projection directions, obtaining the inter-frame correlation coefficient of the horizontal projection and the vertical projection, selecting the region with the largest horizontal projection correlation coefficient to eliminate the jitter in the horizontal direction, selecting the region with the largest vertical projection correlation coefficient to eliminate the jitter in the vertical direction, and thus completing the de-jitter processing of the video sequence.

[0021] Further, the motion object detection for the video sequence after de - jitter processing is specifically as follows: Obtain the first frame image in the video sequence, calculate the background dynamic coefficient, and initialize the background model with the first frame image;

[0022] Obtain the next frame image as the frame to be detected, perform foreground detection on the frame to be detected using the difference method according to the background model, then update the background model with the frame to be detected and use the updated background model to detect the next frame to be detected in the video sequence;

[0023] Repeat the above process until all frames in the current video sequence are detected. If there is a foreground in the frame to be detected, it is determined that a motion object appears.

[0024] Further, the acquisition of the background dynamic coefficient is as follows in formula (1)

[0025]

[0026] where x i is the x - th pixel of the image frame at the i - th time point in the video sequence, F(x i ) is the pixel value of the input image, p k is the sample value of the background model, and N is the number of valid samples;

[0027] The specific operation of initializing the background model with the first frame image is as follows: Randomly sample from the adjacent pixels of x i until the sample set is full, then the initialized background model sample set is obtained. Take the weighted average or normalized value of all pixel values in the sample set as the pixel value of the initialized background model pixel.

[0028] Further, the update of the background model with the frame to be detected specifically includes: Determine whether there is a complex background in the current frame to be detected. If there is no complex background in the frame to be detected, perform a dilation operation on the detection result and then update the background model. Otherwise, first perform an erosion operation on the detection result and then perform a dilation operation to further update the background model.

[0029] The technical solution of the embodiment of the present invention has at least the following advantages and beneficial effects:

[0030] Through this system and method, it is possible to maintain a good motion object recognition rate even when the background environment of the video image changes. At the same time, before the video image motion detection, the video is subjected to de - jitter processing to further improve the motion object recognition rate. After detecting the motion object, the video source where the motion object appears is prompted in the information display module, and relevant prompt messages can also be sent to the monitoring mobile terminal through the network, thereby ensuring that there are no intrusion events and abnormal equipment deformation in the substation. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Schematic flow diagram of the video monitoring method provided by the present invention. Detailed implementation manner

[0032] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Generally, the components of the embodiments of the present invention described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0033] In the existing unattended substation, although the preliminary detection of moving objects can also be carried out through video monitoring, there are still some defects in the existing detection technology. The detection and judgment of moving objects are still not accurate enough, and even misjudgment is likely to occur, resulting in frequent on-site inspections by management personnel, and the effect of saving labor costs is not good. This application hopes to improve the above defects and reduce the management cost of the substation.

[0034] The system embodiment of this application provides a video monitoring system for a substation, including an image acquisition device, a monitoring main end and a monitoring mobile end. The image acquisition device and the monitoring mobile end are both communicatively connected to the monitoring main end; the image acquisition device is used for acquiring environmental and instrument image data and sending it to the monitoring main end;

[0035] The monitoring main end includes a processing and control module and an information display module. The processing and control module includes an image processing unit and an abnormal warning unit. The image processing unit is used for acquiring and processing the data collected by the image acquisition device. Specifically, it acquires the first frame image in the video sequence, calculates the background dynamic coefficient and initializes the background model with it, acquires the next frame image as the detected frame, performs foreground detection on the detected frame according to the background model, then updates the background model through the detected frame and uses the updated background model to detect the next detected frame in the video sequence. Repeat the above process until all frames in the current video sequence are detected. If there is a foreground in the detected frame, it is determined that a moving object appears; the abnormal warning unit is used for giving a warning according to the processing result of the image processing unit. Specifically, when the image processing unit detects a moving object, the abnormal warning unit gives a warning prompt; the information display module is used for displaying the data collected by the image acquisition device and highlighting the data of the detected moving object;

[0036] The monitoring mobile end is used for viewing the images collected by the image acquisition device through the network and prompting the moving object detection information.

[0037] The information display module includes a resource management and control unit, a configuration unit, a video visualization unit, a business unit and a network unit; the resource docking unit is used to obtain the data output by the image acquisition device and the image processing unit and display it in the resource tree; the configuration unit is used to complete the local configuration of the information display module itself; the video visualization unit is used to complete the local video preview and playback operations of the information display module; the business unit includes a business configuration unit and a business control unit, the business configuration unit is used to complete the configuration of the information display module and the association with the decoding channel, the business control unit is used to complete the window opening, segmentation, roaming and wall-mounting processing of the display window; the network unit is used to complete the request and response processing of the message.

[0038] It can be understood that the information display module in the present system embodiment is a TV wall, a preferred system implementation method, in which the TV wall is composed of multiple large-screen displays, each of which can be divided into multiple areas. The business control unit of the TV wall will display the received video resources on the divided large-screen displays.

[0039] When a moving object is detected in a certain video source, although the display presenting the video source is divided into multiple screens, the video source where the moving object is detected is displayed in full screen to highlight it; when moving objects are detected in two video sources in a display, the display is divided into two screens to display the two video sources where the moving object is detected, which can be the upper and lower screens or the left and right screens; when a more extreme situation occurs, that is, when moving objects are detected in the video sources of all split screens in a display, a warning is given by flashing a red light band at the edge of each split screen; it is understandable that the display screen is usually divided into 4 split screens or 9 split screens.

[0040] Although image acquisition devices are fixed in setting, they may still shake in some special cases. When shaking occurs, the impact on the recognition of moving objects is huge, and misjudgment is very likely to occur.

[0041] In order to avoid interference caused by jitter, the processing control module in this embodiment also includes a jitter processing unit, which is used to remove the middle part of the initial frame image of the video sequence after a preset number of pixels are removed from the top, bottom, left and right as a reference frame, and to perform horizontal and vertical projections on the reference frame and the related frames, and then use the Pearson correlation coefficient to describe the correlation between frames, select the area with the largest correlation between the related frame and the reference frame to repair the jitter, and obtain the repaired video sequence.

[0042] Using the system provided in this embodiment, moving objects in the monitoring area of the substation can be detected in real time and accurately, and displayed prominently in the information display device. The information can also be sent to the monitoring mobile terminal through network communication, allowing substation managers to obtain moving object detection information more flexibly.

[0043] The method embodiment of this application provides a video monitoring method for a substation. As Figure 1 shown, the video monitoring method includes:

[0044] Real-time obtain video image data of the monitored area, select a preset number of frames in the preprocessed image as the selected frames, and generate a video sequence with the selected frames in units of frames, that is, evenly obtain a preset number of frames in the preprocessed image as the selected frames at an average time interval or an average frame interval. The preprocessed image is extracted from the video data at a preset time interval or a preset frame interval.

[0045] It can be understood that due to the complexity of the scene, there is a certain delay between the monitoring screen and the real-time scene. This delay is mainly caused by three reasons. One is that the network itself has a certain delay. The second is the selection of the duration of the preprocessed image during moving object detection. The third is the time of moving object detection. Among them, the delays caused by the first and the third are relatively short, usually within one second. Therefore, in order to reduce the time delay with the reality, the preferred duration of the preprocessed image in this application is 2-3 seconds. Since there are more than 20 images per second in high-definition cameras, and in the substation scene, it is mainly to prevent people or animals from invading the substation. Therefore, in order to improve the efficiency of motion detection, this application evenly obtains a preset number of frames in the preprocessed image as the selected frames at an average time interval or an average frame interval, such as taking one frame every 0.1 second as the selected frame or taking one image every 5 frames as the selected frame, and then forming a video sequence from the selected frames.

[0046] Perform anti-jitter processing on the video sequence through the inter-frame correlation coefficient. Specifically, preprocess the video sequence, and use the image obtained by removing a preset number of pixels from the first frame image in the video sequence up, down, left, and right as the reference frame.

[0047] It can be known that when the picture shakes, if the number of pixels of the reference frame is the same as that of other frames, it is difficult to achieve picture alignment without pixel cropping. Therefore, in the method of this application, the reference frame is preferentially cropped so that the pixel matrix of the reference frame is a part of the other frame images, and through subsequent processing, the part of the other frames that is the same as the pixel matrix of the reference frame can be cropped out to achieve picture alignment; generally, the preset number of cropped pixels is 30-60 pixels; after all the frames in the video sequence are cropped, a new video sequence is formed.

[0048] Perform grayscale processing on all frames in the processed video sequence, and then perform local adaptive binarization processing on the grayscale processed image.

[0049] Grayscale processing is a very common type of image processing. Common grayscale processing methods include the component method, the average method, the maximum method, and the weighted average method. Among them, the component method uses the brightness of the three components in a color image as the grayscale values of three grayscale images, and one of the grayscale images can be selected according to application needs. The average method calculates the average of the brightness of the three components in a color image to obtain a grayscale value. The maximum method uses the maximum value of the brightness of the three components in a color image as the grayscale value of the grayscale image. The weighted average method is obtained by weighted averaging the three components with different weights. It can perform weighted averaging on the RGB three components by adjusting the weight threshold according to the application scenario, and then obtain a more reasonable grayscale image.

[0050] Binarization processing sets the value of each pixel point to 0 or 255. After processing, the image has only two grayscale levels, namely black and white. Binarization processing can further reduce the data volume of the image, and then speed up the image processing speed. To obtain better binarization processing results, adaptive binarization processing is generally selected, that is, the threshold corresponding to each pixel is calculated based on adjacent pixels. If the pixel grayscale level of the image is greater than the image calculation threshold at that point, it is marked as the background, otherwise it is the foreground.

[0051] Perform horizontal and vertical projection calculations on the binarized image. Taking the horizontal projection as an example for illustration, the horizontal projection is to sum the columns of the image array. In the binarized image, the object is black and the background is white. Loop to judge whether the pixel value of each column in each row is black, and count the number of all black pixels in that row; the vertical projection is the same; then calculate the quotient of the covariance and standard deviation of different projection directions of each frame and the reference frame to obtain the inter-frame correlation coefficient of the horizontal projection and vertical projection, that is, calculate the Pearson correlation coefficient of different projection directions of each frame and the reference frame, select the region with the largest horizontal projection correlation coefficient to eliminate the jitter in the horizontal direction, and select the region with the largest vertical projection correlation coefficient to eliminate the jitter in the vertical direction, and then complete the de-jitter processing of the video sequence.

[0052] Perform moving object detection on the de-jittered video sequence, that is, obtain the first frame image in the video sequence, calculate the background dynamic coefficient and initialize the background model with the first frame image; among them, the background dynamic coefficient is obtained as shown in the following formula (1)

[0053]

[0054] where, x i is the x-th pixel of the image frame at the i-th time point in the video sequence, F(x i ) is the pixel value of the input image, p k is the sample value of the background model, and N is the number of valid samples.

[0055] The initialization of the background model using the first frame image is specifically as follows: randomly sample from the adjacent pixels of x i until the sample set is full, and then the initialized background model sample set is obtained. The weighted average or normalized value of all pixel values in the sample set is used as the pixel value of the initialized background model pixel.

[0056] For simplicity of operation, the pixels of the first frame can be directly used as the pixels of the initialized background model.

[0057] Obtain the next frame image as the detected frame, and use the difference method to perform foreground detection on the detected frame according to the background model, that is, subtract the detected frame from the background model to obtain the foreground image. In this method, the accurate acquisition of the background model is crucial; therefore, in this method, each detected frame is used to update the background model, and the updated background model is used to detect the next detected frame in the video sequence; by updating the background model frequently, a more accurate background model can be obtained.

[0058] Among them, the update of the background model by the detected frame specifically includes determining whether there is a complex background in the current detected frame. Here, a method of randomly selecting the value of a pixel point in the neighborhood of each pixel point in the image as the background model sample value and randomly updating the model sample value of its neighborhood pixels is used to establish the background model. The selection of the pixel point neighborhood is specifically a circle with a preset radius value centered on the pixel point; we use the background dynamic coefficient to determine whether there is a complex background, and the specific method is as follows in formula (2)

[0059]

[0060] where R is the radius, μ is a fixed coefficient with a value range of 0.3 to 0.6, and τ is the judgment threshold, usually taking values of 4 to 6.

[0061] If there is no complex background in the detected frame, perform a dilation operation on the detection result and then update the background model; otherwise, first perform an erosion operation on the detection result and then perform a dilation operation to update the background model; the updated background model is used to perform the motion detection of the next frame. By continuously updating the background model in real time, the detection of moving objects can be more accurate.

[0062] Repeat the above process until all frames in the current video sequence are detected. If there is a foreground in the detected frame, it is determined that a moving object appears.

[0063] After detecting a moving object in the video, the information display module prompts that there is a moving object in the corresponding video, and at the same time sends the information of the appearance of the moving object to the monitoring mobile terminal.

[0064] Since the video monitoring of the substation is long-term and continuous, if dynamic detection is performed on each segment of the image, the computing load will be very large. In actual situations, the proportion of frames with moving objects is relatively low. To reduce the computing load, the present application also proposes the following method.

[0065] In another method implementation of the present application, based on the previous method embodiment, the time periods with high-frequency occurrence of moving objects are determined through historical data, and potential risk time periods are set up. For example, it is known from historical data that the time periods with high-frequency occurrence of moving objects are from 9 to 11 in the morning and from 13 to 17 in the afternoon, and the potential risk time period is from 1 to 4 in the early morning. Then, selected frames with the same preset value as in the first method embodiment are obtained from each preprocessed image in these time periods; in this embodiment, if it is not within the above time periods, the preset value is adjusted downwards, and the specific value depends on the situation of the substation scene. As can be known from the above example, different preset numbers of selected frames of preprocessed images can be set in different time periods according to the specific situation of the scene to achieve the purpose of reducing computing.

[0066] In addition to the above method, the image frames suspected of having moving objects can be roughly located, and then the above method implementation is used to perform high-accuracy moving object detection on a preset number of consecutive image frames before and after this frame; the rough location can be to extract one frame of image per second, use the image frame extracted in the previous second as the background image, and then subtract the frame extracted in the next second from the frame in the previous second. If the difference exceeds the preset value, it is considered that a moving object has appeared between these two frames of images. At this time, the method of the previous embodiment is used to perform precise moving object detection; through this method, the computing amount can also be greatly reduced, and the computing load of the system can be reduced.

[0067] Through the system and method of the present application, it is possible to maintain a good recognition rate of moving objects even when the background environment of the video image is variable. At the same time, before the moving object detection of the video image, the video is de-jitter processed to further improve the recognition rate of moving objects. After detecting the moving object, the video source where the moving object appears is prompted in the information display module, and relevant prompt messages can also be sent to the monitoring mobile terminal through the network, so as to ensure that there are no intrusion events and abnormal equipment deformation in the substation.

[0068] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A video monitoring system for a substation, characterized in that, It includes an image acquisition device, a monitoring host and a monitoring mobile terminal. The image acquisition device and the monitoring mobile terminal are both communicatively connected to the monitoring host; The image acquisition device is used to collect environmental and instrument image data and send it to the monitoring host; The monitoring host includes a processing and control module and an information display module. The processing and control module includes an image processing unit and an abnormal warning unit. The image processing unit is used to obtain and process the data collected by the image acquisition device. Specifically, it obtains the first frame image in the video sequence, calculates the background dynamic coefficient and uses it to initialize the background model, obtains the next frame image as the detected frame, performs foreground detection on the detected frame according to the background model, then updates the background model through the detected frame and uses the updated background model to detect the next detected frame in the video sequence. Repeat the above process until all frames in the current video sequence are detected; if there is a foreground in the detected frame, it is determined that a moving object appears. The abnormal warning unit is used to give a warning according to the processing result of the image processing unit. Specifically, when the image processing unit detects a moving object, the abnormal warning unit gives a warning prompt; The information display module is used to display the data collected by the image acquisition device and prominently display the data of the detected moving object; Among them, the specific method of initializing the background model with the first frame image is to randomly sample from the adjacent pixels of any pixel in the reference frame until the sample set is full to obtain the initialized background model sample set, and use the weighted average value or normalized value of all pixel values in the sample set as the pixel value of the initialized background model pixel; The update of the background model through the detected frame specifically includes judging whether there is a complex background based on the background dynamic coefficient. If there is no complex background in the detected frame, perform a dilation operation on the detection result and then update the background model. Otherwise, first perform an erosion operation on the detection result and then perform a dilation operation to further update the background model; among them, judging whether there is a complex background based on the background dynamic coefficient, the specific method is as follows formula (2) where R is the radius, μ is a fixed coefficient with a value range of 0.3 to 0.6, τ is a judgment threshold, usually taking a value of 4 to 6, and D(x i ) is the background dynamic coefficient; The monitoring mobile terminal is used to view the images collected by the image acquisition device through the network and prompt the moving object detection information.

2. The video monitoring system for a substation according to claim 1, wherein, The information display module includes a resource management and control unit, a configuration unit, a video visualization unit, a service unit and a network unit; the resource docking unit is used to obtain the data output by the image acquisition device and the image processing unit and display it in the resource tree; The configuration unit is used to complete the local configuration of the information display module itself; the video visualization unit is used to complete the local video preview and playback operations of the information display module; the service unit includes a service configuration unit and a service control unit. The service configuration unit is used to complete the configuration of the information display module and the association with the decoding channel. The service control unit is used to complete the window opening, splitting, roaming and wall mounting processing of the display window; the network unit is used to complete the request and response processing of messages.

3. The video monitoring system for a substation according to claim 1, wherein, The processing control module further includes a jitter processing unit, which is used to take the middle part after removing a preset number of pixels from the top, bottom, left, and right of the initial frame image of the video sequence as a reference frame, project the reference frame and the relevant frames in the horizontal and vertical directions, then describe the inter-frame correlation through the Pearson correlation coefficient, select the region with the largest correlation between the relevant frame and the reference frame to achieve jitter repair, and obtain the repaired video sequence.

4. A video monitoring method for a substation, characterized in that, The video monitoring method includes: Obtaining video image data of the monitored area in real time, selecting a preset number of frames in the preprocessed image as selected frames, and generating a video sequence with the selected frames in units of frames; Performing jitter removal processing on the video sequence through the inter-frame correlation coefficient; Performing moving object detection on the video sequence after jitter removal processing. When a moving object is detected in the video, the information display module is used to prompt that there is a moving object in the corresponding video, and at the same time, the information of the appearance of the moving object is sent to the monitoring mobile terminal; Among them, the performing moving object detection on the video sequence after jitter removal processing specifically includes obtaining the first frame image in the video sequence, calculating the background dynamic coefficient, and initializing the background model with the first frame image; among them, the initializing the background model with the first frame image specifically includes randomly sampling from the adjacent pixels of any pixel of the reference frame until the sample set is filled to obtain the initialized background model sample set, and taking the weighted average or normalized value of all pixel values in the sample set as the pixel value of the initialized background model pixel; Obtaining the next frame image as the detected frame, performing foreground detection on the detected frame using the difference method according to the background model, and then updating the background model with the detected frame and using the updated background model to detect the next detected frame in the video sequence; Repeating the above process until all frames in the current video sequence are detected. If there is a foreground in the detected frame, it is determined that a moving object appears; The updating the background model with the detected frame specifically includes determining whether there is a complex background in the current detected frame. If there is no complex background in the detected frame, perform a dilation operation on the detection result and then update the background model. Otherwise, first perform an erosion operation on the detection result and then perform a dilation operation to update the background model; among them, determining whether there is a complex background in the current detected frame specifically includes: determining whether there is a complex background based on the background dynamic coefficient, and the specific method is as follows formula (2) where R is the radius, μ is a fixed coefficient with a value range of 0.3 to 0.6, τ is a judgment threshold, usually taking a value of 4 to 6, and D(x i ) is the background dynamic coefficient.

5. The video monitoring method for a substation according to claim 4, wherein The selecting a preset number of frames in the preprocessed image as selected frames and generating a video sequence with the selected frames in units of frames specifically includes uniformly obtaining a preset number of frames in the preprocessed image as selected frames at an average time interval or an average frame interval, and the preprocessed image is extracted from the video data at a preset time interval or a preset frame interval.

6. The video monitoring method for a substation according to claim 4, wherein The performing jitter removal processing on the video sequence through the inter-frame correlation coefficient specifically includes preprocessing the video sequence, and taking the image obtained by removing a preset number of pixels from the top, bottom, left, and right of the first frame image in the video sequence as a reference frame; Performing grayscale processing on all frames in the video sequence, and then performing local adaptive binarization processing on the grayscale processed image; Perform horizontal and vertical projection calculations on the binarized image, then calculate the quotient of the covariance and standard deviation of different projection directions for each frame and the reference frame to obtain the inter-frame correlation coefficient of the horizontal and vertical projections. Select the region with the largest horizontal projection correlation coefficient to eliminate jitter in the horizontal direction, and select the region with the largest vertical projection correlation coefficient to eliminate jitter in the vertical direction, thereby completing the anti-jitter processing of the video sequence.

7. The video monitoring method for a substation according to claim 4, characterized in that, The background dynamic coefficient is obtained as shown in the following formula (1) where x i is the x-th pixel of the image frame at the i-th time point in the video sequence, F(x i ) is the pixel value of the input image, p k is the sample value of the background model, and N is the number of valid samples; The specific method for initializing the background model using the first frame image is to randomly sample from the adjacent pixels of x i until the sample set is filled, thus obtaining the initialized background model sample set. The weighted average or normalized value of all pixel values in the sample set is used as the pixel value of the initialized background model pixel.

Images