Method, device, electronic equipment and medium for monitoring and managing confidential archive compact shelving

Abstract

The present invention relates to the field of image processing technology, and provides a method, device, electronic device, and medium for monitoring and managing compact shelving for confidential archives. The method comprises processing a monitoring image to obtain image data; inputting the image data into a material recognition model to obtain material information; using a dynamic Gaussian blur algorithm to perform blurring on areas in the monitoring image whose material information is a preset archive material; using the monitoring image and a motion prediction model to obtain a predicted archive trajectory at at least one future moment, and blurring on areas in the monitoring image whose material information is a preset archive material according to the predicted archive trajectory. The present invention reduces manual intervention links and operational complexity through the process of archive material recognition, motion trajectory prediction, and real-time fuzzy processing, while ensuring the efficiency and consistency of the processing process, avoiding omissions or errors that may be caused by human operation, and improving the security and supervision efficiency of confidential archive management.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method, device, electronic equipment and medium for monitoring and managing compact shelving for confidential archives. Background Art

[0002] The safekeeping of confidential archives is a crucial component of national security and information confidentiality, requiring strict safeguards against information leaks. Traditional confidential archive storage typically relies on manual oversight or basic monitoring equipment, but these methods present significant drawbacks. Manual oversight is inefficient and prone to omissions, making it difficult to track archive access in real time. While basic monitoring equipment can record access, the camera may capture key information on the archive box, such as the file number and title. This information is inherently confidential, potentially leading to secondary leaks and creating new security risks.

[0003] Currently, video masking technology can be used to prevent sensitive areas from being recorded by video surveillance. However, this technology requires pre-setting fixed masking areas. The location of files in the archive room changes dynamically due to daily use, resulting in a mismatch between the pre-set masking areas and the actual sensitive locations, making it difficult to effectively protect information in a sustained manner. Alternatively, OCR technology can be combined with privacy coding to blur text after recognition. However, this method is mainly designed for flat text and requires the recognized text to remain horizontal. However, in actual archive room environments, file boxes are often tilted, or the images captured by the surveillance are tilted due to the angle of the surveillance installation. This makes it very easy for the OCR algorithm to fail to detect text. In complex scenarios, sensitive information can be missed due to the inability to detect text, and the problem of information leakage is not fundamentally solved. Summary of the Invention

[0004] The problem solved by the present invention is how to improve the security and supervision efficiency of archive management.

[0005] In order to solve the above problems, the present invention provides a method, device, electronic equipment and medium for monitoring and managing confidential archive compact shelving.

[0006] In a first aspect, the present invention provides a method for monitoring and managing a compact storage rack for confidential archives, comprising:

[0007] Processing the surveillance images obtained by shooting the compact rack of confidential archives to obtain image data;

[0008] Inputting the image data into a material recognition model to obtain material information, wherein the material recognition model is obtained by marking the material of the historical image data of the confidential archives compact shelving and training based on the marking results and the historical image data;

[0009] Using a dynamic Gaussian blur algorithm to perform blur processing on the area of ​​the monitoring image where the material information is the preset file material;

[0010] The monitoring image and the motion prediction model are used to obtain a predicted archive trajectory for at least one future moment, and the area in the monitoring image for at least one future moment where the material information is a preset archive material is blurred according to the predicted archive trajectory, wherein the motion prediction model is trained based on historical archive motion information.

[0011] Optionally, the material recognition model includes:

[0012] A feature extraction module, configured to extract features from the image data;

[0013] an attention mechanism module, configured to determine key areas in the image data based on the extracted features;

[0014] A feature fusion module, configured to fuse the extracted features of the key area to obtain a fused feature;

[0015] A classifier is used to classify materials according to the fusion features to obtain material information of the key area.

[0016] Optionally, the step of processing the surveillance image obtained by photographing the confidential archives compact rack to obtain image data includes:

[0017] Obtaining RGB image data and multispectral imaging data according to the monitoring image;

[0018] Grayscale processing, median filtering and histogram equalization are performed on the RGB image data and the multispectral imaging data respectively to obtain the image data.

[0019] Optionally, the blurring of the area in the surveillance image where the material information is a preset file material using a dynamic Gaussian blur algorithm includes:

[0020] obtaining environmental parameters according to the image data;

[0021] The blur radius of the area where the material information is the preset file material is adjusted according to the environmental parameters, and the blur intensity is adjusted according to the blur radius.

[0022] Optionally, the environmental parameter includes light intensity data; and adjusting the blur intensity according to the blur radius includes:

[0023] When the blur radius is less than or equal to the preset radius, the blur intensity is the initial intensity;

[0024] When the blur radius is greater than the preset radius, increasing the blur intensity based on the initial intensity, the blur intensity and the blur radius are positively correlated;

[0025] Alternatively, when the blur radius is greater than the preset radius, an intensity base value is obtained based on a preset radius factor, the initial intensity and the blur radius, an intensity adjustment value is determined according to a preset light intensity factor and the light intensity data, and the blur intensity is obtained by obtaining the sum of the intensity base value and the intensity adjustment value.

[0026] Optionally, the motion prediction model includes:

[0027] an optical flow method module, configured to obtain optical flow information of an area in the surveillance image whose material information corresponds to a preset archival material, and obtain a predicted motion trend of the area whose material information corresponds to the preset archival material at at least one future moment based on the optical flow information;

[0028] A Kalman filter module is used to obtain the archival motion state of the area in the surveillance image where the material information is the preset archival material, and obtain the predicted archival trajectory of the area where the material information is the preset archival material at at least one future moment based on the archival motion state and the predicted motion trend.

[0029] Optionally, the material recognition model further includes an environment fusion module, the output of which is connected to the input of the feature extraction module; the environment fusion module is used to fuse the environment parameters pre-obtained based on the image data with the image data.

[0030] In a second aspect, the present invention provides a monitoring and management device for a compact rack of confidential archives, comprising:

[0031] A processing unit, configured to process the surveillance image obtained by photographing the confidential archives compact rack to obtain image data;

[0032] an identification unit, configured to input the image data into a material identification model to obtain material information, wherein the material identification model is obtained by marking the material of the historical image data of the confidential archives compact shelving and training based on the marking results and the historical image data;

[0033] a blurring unit, configured to use a dynamic Gaussian blurring algorithm to perform blurring processing on the area of ​​the surveillance image where the material information is a preset archival material;

[0034] A prediction unit is configured to utilize the surveillance image and a motion prediction model to obtain a predicted archive trajectory for at least one future moment, and to perform blurring on an area in the surveillance image for at least one future moment in which the material information is a preset archive material according to the predicted archive trajectory, wherein the motion prediction model is trained based on historical archive motion information.

[0035] In a third aspect, the present invention provides an electronic device comprising a memory and a processor;

[0036] The memory is used to store computer programs;

[0037] The processor is used to implement the confidential archive compact shelving monitoring and management method as described in the first aspect when executing the computer program.

[0038] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for monitoring and managing compact shelving of confidential archives as described in the first aspect is implemented.

[0039] The beneficial effect of the confidential archive compact shelving monitoring and management method of the present invention is: the image data obtained from the monitoring image obtained by shooting the confidential archive compact shelving is input into a pre-trained material recognition model to obtain material information, and then a dynamic Gaussian blur algorithm is used to blur the area in the monitoring image whose material information is a preset archive material. Compared with the related art of first identifying text and then performing privacy encoding on the text part, the protection object is upgraded from the text level to the material level, and the blurring is performed after the preset archive material is identified, so that only non-confidential archive access behavior images are retained in the monitoring record, ensuring that confidential content (such as text information after text recognition) is not retained in the monitoring link, eliminating safety hazards from the source, and at the same time ensuring that the closed-loop operation from detection to protection is completed in a short time, increasing the security of archive management and the efficiency of supervision. Surveillance images and a motion prediction model are used to obtain a predicted archive trajectory for at least one future moment. Areas in the surveillance images at at least one future moment whose material information corresponds to a preset archive material are blurred according to the predicted archive trajectory. To address the dynamic changes caused by the movement of the archive box, the motion prediction model is used to predict the archive's movement trajectory in advance, achieving accurate trajectory prediction in a high-frame-rate video stream. Blurring is then performed on areas corresponding to the preset archive material, thereby enhancing the intelligent and automated level of confidential archive management. This embodiment, through the process of archive material identification, motion trajectory prediction, and real-time fuzzification, reduces manual intervention and operational complexity, while ensuring the efficiency and consistency of the processing process, avoiding omissions or errors that may result from human operation, and significantly improving the security and regulatory efficiency of confidential archive management. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flowchart of a method for monitoring and managing a compact rack of confidential archives according to an embodiment of the present invention is provided;

[0041] Figure 2 A structural diagram of a material recognition model according to an embodiment of the present invention;

[0042] Figure 3 Schematic diagram of the structure of a motion prediction model according to an embodiment of the present invention;

[0043] Figure 4 This is a schematic diagram of the structure of a monitoring and management device for a compact rack of confidential files according to an embodiment of the present invention;

[0044] Figure 5 Schematic diagram of the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0045] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings. Although certain embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as being limited to the embodiments described herein. Instead, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0046] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0047] The term "including" and its variations used in this document are open inclusions, that is, "including but not limited to"; the term "based on" means "based at least in part on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "optional embodiments". The relevant definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc. mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0048] It should be noted that the modifications of "one" and "multiple" mentioned in the present invention are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0049] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0050] In response to the problems existing in the above-mentioned related technologies, this embodiment provides a monitoring and management method, device, electronic equipment and medium for a compact shelving system for confidential archives.

[0051] like Figure 1 As shown, an embodiment of the present invention provides a method for monitoring and managing a compact storage rack for confidential files, comprising:

[0052] Step S1: Process the surveillance image obtained by shooting the confidential archive compact rack to obtain image data.

[0053] Specifically, surveillance images are acquired based on surveillance video captured of a compact shelving system containing confidential archives. These surveillance images can be each frame of the surveillance video. These surveillance images are then preprocessed, for example, using bilinear interpolation, bicubic interpolation, or other methods to improve image resolution, and using Gaussian filtering or bilateral filtering for denoising. This improves the quality of the acquired surveillance image data and makes it more suitable for subsequent analysis and processing. The image data can be RGB image data and / or multispectral imaging data.

[0054] Step S2: input the image data into a material recognition model to obtain material information, wherein the material recognition model is obtained by marking the material of the historical image data of the confidential archives compact shelving and training based on the marking results and the historical image data.

[0055] Specifically, the system identifies the materials at various locations in surveillance videos of classified archive shelving, identifies the materials of the archive boxes and their corresponding locations, and subsequently accurately processes the locations of the archive boxes. Historical image data of the classified archive shelving is retrieved in advance, and the material information in the historical image data is labeled. Labeling results corresponding to the historical image data are obtained. A pre-built model is trained using the historical image data and the labeled results as training and test sets, resulting in a trained material recognition model. The material recognition model takes image data as input and outputs material information for each region in the image data. The pre-built model can be constructed using, for example, a convolutional neural network (CNN). Finally, the acquired image data is fed into the trained material recognition model to obtain the material information.

[0056] Step S3: Using a dynamic Gaussian blur algorithm to perform blur processing on the area in the monitoring image where the material information is the preset file material.

[0057] Specifically, based on the material information obtained in step S1, regions corresponding to different material information that are identical or similar to the preset archival material are obtained, and a dynamic Gaussian blur algorithm is used to blur these regions. Algorithms such as Euclidean distance, cosine similarity, KL divergence, and JS divergence can be used to calculate the similarity between the different material information and the preset archival material. Regions with similarity greater than a preset threshold are then identified as target archival regions, i.e., regions requiring privacy processing, effectively increasing the accuracy of archival privacy processing.

[0058] Step S4: Utilize the surveillance image and the motion prediction model to obtain a predicted archive trajectory for at least one future moment, and perform blurring on the area in the surveillance image for at least one future moment where the material information is a preset archive material according to the predicted archive trajectory, wherein the motion prediction model is trained based on historical archive motion information.

[0059] Specifically, because the monitoring and management is based on real-time surveillance of the archive room, the location of the archives will change in real time when, for example, staff members retrieve archives. Therefore, a motion prediction model is used to predict the archive's motion position at at least one future moment. Based on the predicted archive trajectory, regions in the surveillance image at at least one future moment with material information corresponding to a preset archive material are blurred in real time, thereby enhancing the security of archive management. First, historical archive motion information from historical surveillance videos is retrieved to train a preset prediction model, resulting in a trained running prediction model. The running prediction model takes the surveillance image as input and outputs a predicted archive trajectory, i.e., the predicted archive motion trajectory. The captured surveillance image is then input into the trained running prediction model to obtain a predicted archive trajectory. Based on the predicted archive trajectory, the location of regions with material information corresponding to the preset archive material in the surveillance image at at least one future moment is determined, and the blurring is performed using a dynamic Gaussian blur algorithm. For example, surveillance images are captured in real time. When a surveillance image is acquired at a future moment, the location of regions with material information corresponding to the preset archive material, as determined by the predicted archive trajectory, in the surveillance image at that moment is blurred.

[0060] In addition, in the above process, data of the time and operation of each process is recorded and transmitted to the back-end server in a wired manner through the Ethernet interface of the camera communication unit. After receiving the data, the back-end server stores it in a secure database using encryption technology. At the same time, the stored data can be managed and analyzed, such as querying and retrieving according to conditions such as time and file number, providing support for file management, facilitating subsequent file management investigations and other operations, and helping to achieve the organic unity of the accuracy of confidential file supervision and privacy protection. While blurring confidential information, it fully records key behavioral data such as the time, personnel, and location of file access, which not only meets the traceability needs of the file circulation process in security management, but also avoids privacy risks caused by excessive collection of sensitive information.

[0061] Furthermore, in terms of hardware design, smart cameras with open interfaces (such as Huawei's SDC6582-I) were selected. Leveraging the computing power and interfaces provided by the cameras, the monitoring and management method for confidential archives compact shelving systems was embedded. The camera's hardware architecture primarily consists of an image acquisition unit, an algorithm processing unit, a storage unit, and a communication unit. The image acquisition unit consists of a high-definition lens and image sensor. The high-definition lens offers high resolution, a wide viewing angle, and excellent optical performance, enabling clear capture of the surface information of the archive boxes under varying lighting conditions. The camera's built-in high-performance chip, with its multi-core processing capabilities and powerful floating-point computing capabilities, provides the computing resources necessary for the monitoring and management method for confidential archives compact shelving systems. High-speed storage media is used for temporary storage of captured image data and intermediate results during processing, ensuring fast data read and write speeds and avoiding processing delays. The communication unit supports multiple wired communication interfaces, such as Ethernet, enabling stable and high-speed transmission of processed privacy-encoded data to the backend server.

[0062] In this embodiment, the image data obtained from the surveillance image obtained by shooting the confidential archive compact shelving is input into a pre-trained material recognition model to obtain material information, and then the dynamic Gaussian blur algorithm is used to blur the area in the surveillance image whose material information is the preset archive material. Compared with the related art of first identifying the text and then performing privacy encoding on the text part, the protection object is upgraded from the text level to the material level, and the blurring is performed after the preset archive material is identified, so that only non-confidential archive access behavior images are retained in the surveillance record, ensuring that confidential content (such as text information after text recognition) is not retained in the monitoring link, eliminating security risks from the source, and at the same time ensuring that the closed-loop operation from detection to protection is completed in a short time, thereby increasing the security of archive management and the efficiency of supervision. Surveillance images and a motion prediction model are used to obtain a predicted archive trajectory for at least one future moment. Areas in the surveillance images at at least one future moment whose material information corresponds to a preset archive material are blurred according to the predicted archive trajectory. To address the dynamic changes caused by the movement of the archive box, the motion prediction model is used to predict the archive's movement trajectory in advance, achieving accurate trajectory prediction in a high-frame-rate video stream. Blurring is then performed on areas corresponding to the preset archive material, thereby enhancing the intelligent and automated level of confidential archive management. This embodiment, through the process of archive material identification, motion trajectory prediction, and real-time fuzzification, reduces manual intervention and operational complexity, while ensuring the efficiency and consistency of the processing process, avoiding omissions or errors that may result from human operation, and significantly improving the security and regulatory efficiency of confidential archive management.

[0063] Optionally, the material recognition model includes:

[0064] A feature extraction module, configured to extract features from the image data;

[0065] an attention mechanism module, configured to determine key areas in the image data based on the extracted features;

[0066] A feature fusion module, configured to fuse the extracted features of the key area to obtain a fused feature;

[0067] A classifier is used to classify materials according to the fusion features to obtain material information of the key area.

[0068] Specifically, if Figure 2As shown in the figure, the material recognition model includes an adjusted feature extraction module, an attention mechanism module, a feature fusion module, and a classifier. The feature extraction module can include convolutional neural networks such as ResNet and VGG. Taking ResNet as an example, it consists of multiple convolutional layers, residual blocks, and pooling layers. These layers can gradually extract low-level features (such as edges and corners) and high-level features (such as texture and shape) from the image. If the image data is multispectral imaging data, adjustments can be made based on the characteristics of the multispectral data. For example, the number of input channels must match the number of channels in the multispectral data to extract more accurate features. The attention mechanism module can include a Squeeze-and-Excitation (SE) module or a Convolutional Block Attention Module (CBAM). Taking the SE module as an example, it first performs global average pooling on the features extracted from the image data, compressing them into a one-dimensional vector. Then, it passes through two fully-connected layers: the first reduces the vector dimension, and the second restores it to its original number of channels. A sigmoid function is used to map the output values ​​to the range [0, 1] to obtain an attention weight for each channel. Finally, the attention weights are multiplied channel-by-channel with the corresponding feature map to enhance the feature representation of important channels. The feature fusion module fuses the image features processed by the attention mechanism. A simple concatenation method can be used to concatenate multiple features along the channel dimension to produce a fused feature map. The classifier performs global average pooling on the fused feature map, converting it into a one-dimensional vector. The fused feature map then passes through one or more fully-connected layers, followed by a softmax layer, to output a probability distribution corresponding to each material type, i.e., a confidence score.

[0069] Optionally, the step of processing the surveillance image obtained by photographing the confidential archives compact rack to obtain image data includes:

[0070] Obtaining RGB image data and multispectral imaging data according to the monitoring image;

[0071] Grayscale processing, median filtering and histogram equalization are performed on the RGB image data and the multispectral imaging data respectively to obtain the image data.

[0072] Specifically, in an archive room, the position of a surveillance camera is usually fixed. The captured images may have problems such as uneven lighting and noise interference due to the placement of the archives and changes in lighting. Therefore, the captured surveillance images need to be preprocessed to improve the accuracy of image data acquisition. This application obtains RGB image data and multispectral imaging data based on the surveillance images, and then grayscales the two types of data separately, converting the color image into a grayscale image to reduce the amount of data and highlight the material characteristics. Then, a filtering operation such as median filtering is performed to remove noise. Finally, histogram equalization is performed to enhance the image contrast, making the paper fiber texture on the surface of the archive box clearer, providing good input for material recognition.

[0073] Optionally, the blurring of the area in the surveillance image where the material information is a preset file material using a dynamic Gaussian blur algorithm includes:

[0074] obtaining environmental parameters according to the image data;

[0075] The blur radius of the area where the material information is the preset file material is adjusted according to the environmental parameters, and the blur intensity is adjusted according to the blur radius.

[0076] Specifically, environmental parameters can include light intensity, backlight detection, and light tilt angle. The light sensor built into the surveillance camera can be used to obtain raw data on ambient light intensity. The range is 0-2000 Lux, with larger values ​​indicating stronger light. Since the sensor typically outputs an analog signal, and computers can only process digital signals, analog-to-digital conversion is required to convert the analog light intensity data into digital form. The converted digital data is mapped to the interval [0, 1] to obtain the light intensity factor I. When I = 0, the environment is completely dark, and when I = 1, the ambient light intensity reaches a strong 2000 Lux. Normalization facilitates the unified calculation of subsequent algorithms. For backlit scene detection, the image analysis module is used to calculate the gradient of RGB image data. The gradient can reflect the rate of change of pixel values ​​in the image and can help identify information such as edges and textures in the image. Generally, the RGB image is converted into a grayscale image first, and then the gradient is calculated using methods such as the Sobel operator. The screening conditions are set, such as the pixel gradient amplitude is greater than 50 and the brightness value exceeds 200 gray levels. The number of pixels that meet these conditions is counted and their proportion in the total number of pixels in the entire image is calculated. When the proportion of pixels that meet the conditions exceeds 15%, the area is determined to be a backlit area and a binary label δ is generated. 逆光, where 1 represents backlighting and 0 represents normal. In the calculation of the tilt angle, the optical flow method is used to track the motion trajectory of at least 10 feature corner points on the file box, and the three-dimensional perspective transformation matrix is ​​used to fit the spatial posture of the edge contour of the file box through the least squares method. The three-dimensional perspective transformation matrix can describe the projection relationship of the object in three-dimensional space. The least squares method is used to find the optimal fitting parameters, and finally the tilt angle θ around the optical axis of the camera is solved, where 0° represents the front view plane and 90° represents the vertical side. The environmental parameters are passed as input to the dynamic Gaussian blur algorithm to obtain the blur radius. The blur radius R uses the basic value R0=8 pixels, and is dynamically adjusted in combination with the illumination compensation item and the tilt enhancement item. The specific formula is R=R0×(1+0.3I+0.5δ 逆光 +0.01θ), which can adaptively adjust the blur effect according to different environmental parameters.

[0077] It should be noted that the blur intensity can also be adjusted according to the blur radius.

[0078] Optionally, the environmental parameter includes light intensity data; and adjusting the blur intensity according to the blur radius includes:

[0079] When the blur radius is less than or equal to the preset radius, the blur intensity is the initial intensity;

[0080] When the blur radius is larger than the preset radius, the blur intensity is increased based on the initial intensity. The blur intensity is positively correlated with the blur radius. For example, when the blur radius R increases, the blur intensity increases accordingly. A functional relationship can be set, such as blur intensity S=kRS 初 , where k represents the preset radius factor, and the appropriate value is determined through experiments.

[0081] It should be noted that from the perspective of human visual perception, a larger blur radius will make the details and edges in the image more blurred and unclear, which is consistent with our intuitive understanding of blur intensity. In practical applications, such as when blurring a privacy area, we hope to increase the blur intensity by increasing the blur radius to better hide sensitive information. If the blur radius increases but the blur intensity does not increase accordingly, the expected blur effect cannot be achieved, and privacy protection or other purposes that require blurring cannot be effectively achieved. Therefore, the blur intensity is set to be positively correlated with the blur radius. In order to avoid the leakage of archival information due to the small blur intensity resulting from a small blur radius, the blur intensity is set to the initial intensity when the blur radius is less than or equal to the preset radius. The preset radius can be set according to the size of the archival text.

[0082] Alternatively, when the blur radius is greater than the preset radius, an intensity base value is obtained based on a preset radius factor, the initial intensity and the blur radius, an intensity adjustment value is determined according to a preset light intensity factor and the light intensity data, and the blur intensity is obtained by obtaining the sum of the intensity base value and the intensity adjustment value.

[0083] Specifically, in addition to the influence of radius, light intensity can also make additional adjustments to the blur intensity. In a strong light environment (the light intensity factor I is large), the blur intensity can be appropriately increased to better deal with possible reflections, overexposure and other problems; in a weak light environment, the blur intensity is appropriately reduced to prevent the image from becoming too blurred and losing details. Since the blur intensity does not change when the blur radius is less than or equal to the preset radius, this method is only performed on the basis that the blur radius is greater than the preset radius. First, based on the positive correlation between the blur intensity and the blur radius, the intensity base value is obtained, and then the intensity adjustment value for the light intensity is determined based on the preset light intensity factor and the light intensity data. The sum of the intensity base value and the intensity adjustment value is taken to obtain the blur intensity, such as blur intensity S=kRS 初 +dIS 初 , where d represents the preset light intensity factor.

[0084] In this embodiment, the blur intensity is dynamically adjusted by adjusting the blur radius and illumination intensity. In scenarios involving privacy protection, such as blurring sensitive information such as faces and license plates in surveillance video, a larger blur radius means wider coverage, allowing more surrounding areas to be included in the blur range. Increasing the blur intensity accordingly can more effectively render details in sensitive areas completely unrecognizable, thereby better protecting personal privacy and sensitive information and preventing information leakage. Different application scenarios have different requirements for the degree of blur. By adjusting the blur intensity based on the blur radius, the needs of various scenarios can be flexibly met. In practical applications, image content and characteristics vary widely. Adjusting the blur intensity based on the blur radius gives the dynamic Gaussian blur algorithm greater adaptability. For example, for areas with complex textures, a larger blur radius and correspondingly higher blur intensity may be required to achieve the desired blur effect; for areas with simpler textures, a smaller blur radius and lower blur intensity can be used. This automatic adjustment of the blur intensity based on the specific image conditions eliminates the need for frequent manual adjustment of multiple parameters, improving the algorithm's versatility and practicality.

[0085] Alternatively, as Figure 3 As shown, the motion prediction model includes an optical flow module and a Kalman filter module;

[0086] an optical flow method module, configured to obtain optical flow information of an area in the surveillance image whose material information corresponds to a preset archival material, and obtain a predicted motion trend of the area whose material information corresponds to the preset archival material at at least one future moment based on the optical flow information;

[0087] A Kalman filter module is used to obtain the archival motion state of the area in the surveillance image where the material information is the preset archival material, and obtain the predicted archival trajectory of the area where the material information is the preset archival material at at least one future moment based on the archival motion state and the predicted motion trend.

[0088] Specifically, the optical flow module first extracts feature points from continuous frame surveillance images in the surveillance video. Common feature point extraction algorithms include SIFT (Scale-Invariant Feature Transform), SURF (Speeded Robust Features) or ORB (Oriented Rotation Brief). Feature points can represent key positions in the image and have a certain degree of stability and uniqueness. Based on the extracted feature points, an optical flow algorithm (such as the Lucas-Kanade optical flow algorithm and the Farneback optical flow algorithm) is used to calculate the motion vectors of feature points between adjacent frames. The optical flow represents the movement direction and speed of objects in the image between consecutive frames. The preliminary motion information of the object can be obtained through optical flow calculation. The Kalman filter module is used to predict the archive's motion trajectory. It first constructs the state transition matrix, process noise covariance matrix, measurement matrix, and measurement noise covariance matrix. The Kalman filter formula then calculates the Kalman gain, K, to balance the weights of the predicted and measured values. The state transition matrix is ​​used to predict the current state, yielding a predicted state value for at least one future moment. The motion vector calculated using the optical flow method is then used as the measurement value (converting the optical flow vector into position information). The predicted value is then updated using the Kalman gain to obtain a more accurate state estimate. This process is repeated to predict the state of the archive box for at least one future moment, resulting in a pre-movement offset (Δx, Δy), effectively the predicted archive trajectory for at least one future moment.

[0089] Furthermore, after obtaining the predicted archive trajectory for at least one future moment, a dynamic Gaussian blur algorithm or a dynamic occlusion mask generation module is used to blur the area in the surveillance image at at least one future moment where the material information corresponds to the preset archive material. The dynamic occlusion mask generation module is used to form a blurred area synchronized with the archive box's movement at the effective moment. Based on the binary region mask output by the material recognition module, a morphological dilation operation is performed. A 5x5 structuring element (e.g., a square structuring element) is selected and dilated by 5 pixels to compensate for any boundary errors in material recognition and ensure complete coverage of the archive box area. The expanded mask is then translated in real time on the image plane based on the pre-movement offset (Δx, Δy) obtained by the Kalman filter module, synchronizing its position with the predicted archive box's movement trajectory, thereby forming a blurred area synchronized with the archive box's movement.

[0090] Optionally, the material recognition model further includes an environment fusion module, the output of which is connected to the input of the feature extraction module; the environment fusion module is used to fuse the environment parameters pre-obtained based on the image data with the image data.

[0091] Specifically, the environmental fusion module precedes the feature extraction module. It first acquires and processes environmental parameters using various sensors or devices, such as light intensity from a photosensor and ambient temperature from a temperature sensor. Indirect environmental parameters, such as those from the camera, can be obtained from image metadata or by referring to the environmental parameter acquisition methods described above. The acquired environmental parameters undergo preprocessing, including data cleaning (removing outliers and missing values), normalization (mapping parameter values ​​of varying ranges to a uniform interval, such as [0, 1]), and format conversion (converting the parameters to a format suitable for image data fusion). This ensures that the environmental parameters are accurately incorporated into the subsequent fusion process. Next, the input image data (RGB or multispectral imaging data) undergoes conventional preprocessing, such as resizing the image to meet the requirements of the feature extraction module and normalizing the image pixel values ​​(scaling the pixel value range from [0, 255] to [0, 1]). The environmental parameters are then fused with the image data. The preprocessed environmental parameters are concatenated with the image data as additional channels or dimensions. For example, if the environmental parameter is a three-dimensional vector containing light intensity, temperature, and humidity, and the RGB image data has three channels (red, green, and blue), the environmental parameter vector can be concatenated with the RGB image data along the channel dimension to form a new data set with six channels. A weight is assigned to each environmental parameter based on its influence on the image features. The environmental parameters are then weighted and summed with the image data. For example, if light intensity has a greater influence on image features, a higher weight can be assigned to light intensity, allowing it to more significantly influence the image data during the fusion process. The environmental parameters are encoded as feature vectors and then embedded into the feature space of the image data. For example, a neural network can be used to map the environmental parameters to a low-dimensional feature vector. This feature vector is then convolved with the image data to obtain a feature map, and then element-wise addition or other fusion operations are performed. The fused environmental parameters and image data are then output to the feature extraction module as input data. The feature extraction module will perform subsequent feature extraction operations based on these fused data, so that the model can learn the relationship between environmental parameters and image features, thereby improving the accuracy and robustness of material recognition.

[0092] In this embodiment, the light intensity information may affect the direction and focus of image feature extraction in subsequent convolution operations. An environment fusion module is set before the feature extraction module to interact the light intensity information with the image data in the early stage of model processing. If the light intensity is high, some image details may be lost due to excessive brightness. After the light intensity information is fused early, the model can adjust the parameters of the convolution kernel or the convolution operation mode according to the light intensity, so that the extracted image features are more in line with the actual situation under the current light intensity conditions, reducing the risk of information loss, and making the model processing process closer to the real situation, which is more conducive to the subsequent accurate judgment of the object material.

[0093] like Figure 4 As shown, an embodiment of the present invention provides a confidential file compact rack monitoring and management device 400, comprising:

[0094] The processing unit 410 is used to process the surveillance image obtained by photographing the confidential archives compact rack to obtain image data;

[0095] The recognition unit 420 is configured to input the image data into a material recognition model to obtain material information, wherein the material recognition model is obtained by marking the material of the historical image data of the confidential archives compact shelving and training the model based on the marking results and the historical image data;

[0096] A blurring unit 430 is configured to use a dynamic Gaussian blurring algorithm to blur the area of ​​the surveillance image where the material information is a preset archival material;

[0097] A prediction unit 440 is configured to utilize the surveillance image and a motion prediction model to obtain a predicted archive trajectory for at least one future moment, and to perform blurring on an area in the surveillance image for at least one future moment in which the material information is a preset archive material according to the predicted archive trajectory, wherein the motion prediction model is trained based on historical archive motion information.

[0098] like Figure 5 As shown, an electronic device 500 provided by an embodiment of the present invention includes a memory 510 and a processor 520; the memory 510 is used to store computer programs; the processor 520 is used to implement the above-mentioned confidential archive compact shelving monitoring and management method when executing the computer program.

[0099] In other words, an electronic device 500 includes a memory 510 and a processor 520 coupled to the memory 510; the memory 510 is configured to store a computer program; and the processor 520 is configured to perform the following operations when executing the computer program:

[0100] Processing the surveillance images obtained by shooting the compact rack of confidential archives to obtain image data;

[0101] Inputting the image data into a material recognition model to obtain material information, wherein the material recognition model is obtained by marking the material of the historical image data of the confidential archives compact shelving and training based on the marking results and the historical image data;

[0102] Using a dynamic Gaussian blur algorithm to perform blur processing on the area of ​​the monitoring image where the material information is the preset file material;

[0103] The monitoring image and the motion prediction model are used to obtain a predicted archive trajectory for at least one future moment, and the area in the monitoring image for at least one future moment where the material information is a preset archive material is blurred according to the predicted archive trajectory, wherein the motion prediction model is trained based on historical archive motion information.

[0104] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the above-mentioned method for monitoring and managing compact shelving of confidential archives is implemented.

[0105] In other words, a non-volatile computer-readable storage medium stores a computer program, which, when executed by a processor, causes the processor to perform the following operations:

[0106] Processing the surveillance images obtained by shooting the compact rack of confidential archives to obtain image data;

[0107] Inputting the image data into a material recognition model to obtain material information, wherein the material recognition model is obtained by marking the material of the historical image data of the confidential archives compact shelving and training based on the marking results and the historical image data;

[0108] Using a dynamic Gaussian blur algorithm to perform blur processing on the area of ​​the monitoring image where the material information is the preset file material;

[0109] The monitoring image and the motion prediction model are used to obtain a predicted archive trajectory for at least one future moment, and the area in the monitoring image for at least one future moment where the material information is a preset archive material is blurred according to the predicted archive trajectory, wherein the motion prediction model is trained based on historical archive motion information.

[0110] An electronic device 500 that can serve as a server or client of the present invention will now be described, which is an example of a hardware device that can be applied to various aspects of the present invention. The electronic device 500 is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device 500 can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0111] Electronic device 500 includes a computing unit that can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) or loaded from a storage unit into a random access memory (RAM). The RAM can also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. An input / output (I / O) interface is also connected to the bus.

[0112] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM). In this application, the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network elements. Some or all of these units can be selected based on actual needs to achieve the objectives of the embodiments of the present invention. Furthermore, the functional units in the various embodiments of the present invention can be integrated into a single processing unit, each unit can exist physically separately, or two or more units can be integrated into a single unit. These integrated units can be implemented in either hardware or software functional units.

[0113] Although the present invention is disclosed as above, the protection scope of the present invention is not limited thereto. Those skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will fall within the protection scope of the present invention.

Claims

1. A method for monitoring and managing a compact rack of confidential archives, characterized in that: include: Processing the surveillance images obtained by shooting the compact rack of confidential archives to obtain image data; Inputting the image data into a material recognition model to obtain material information, wherein the material recognition model is obtained by marking the material of the historical image data of the confidential archives compact shelving and training based on the marking results and the historical image data; Using a dynamic Gaussian blur algorithm to perform blur processing on the area of ​​the monitoring image where the material information is the preset file material; The monitoring image and the motion prediction model are used to obtain a predicted archive trajectory for at least one future moment, and the area in the monitoring image for at least one future moment where the material information is a preset archive material is blurred according to the predicted archive trajectory, wherein the motion prediction model is trained based on historical archive motion information.

2. The method for monitoring and managing confidential archives compact racks according to claim 1, characterized in that: The material recognition model includes: A feature extraction module, configured to extract features from the image data; an attention mechanism module, configured to determine key areas in the image data based on the extracted features; A feature fusion module, configured to fuse the extracted features of the key area to obtain a fused feature; A classifier is used to classify materials according to the fusion features to obtain material information of the key area.

3. The method for monitoring and managing confidential archives compact shelving according to claim 1, characterized in that: The image data obtained by processing the surveillance image obtained by photographing the confidential archives compact rack includes: Obtaining RGB image data and multispectral imaging data according to the monitoring image; Grayscale processing, median filtering and histogram equalization are performed on the RGB image data and the multispectral imaging data respectively to obtain the image data.

4. The method for monitoring and managing confidential archives compact shelving according to claim 1, characterized in that: The blurring of the area in the surveillance image where the material information is the preset file material using the dynamic Gaussian blur algorithm includes: obtaining environmental parameters according to the image data; The blur radius of the area where the material information is the preset file material is adjusted according to the environmental parameters, and the blur intensity is adjusted according to the blur radius.

5. The method for monitoring and managing confidential archives compact racks according to claim 4, characterized in that: The environmental parameter includes light intensity data; and adjusting the blur intensity according to the blur radius includes: When the blur radius is less than or equal to the preset radius, the blur intensity is the initial intensity; When the blur radius is greater than the preset radius, increasing the blur intensity based on the initial intensity, the blur intensity and the blur radius are positively correlated; Alternatively, when the blur radius is greater than the preset radius, an intensity base value is obtained based on a preset radius factor, the initial intensity and the blur radius, an intensity adjustment value is determined according to a preset light intensity factor and the light intensity data, and the blur intensity is obtained by obtaining the sum of the intensity base value and the intensity adjustment value.

6. The method for monitoring and managing confidential archives compact shelving according to claim 1, characterized in that: The motion prediction model includes: an optical flow method module, configured to obtain optical flow information of an area in the surveillance image whose material information corresponds to a preset archival material, and obtain a predicted motion trend of the area whose material information corresponds to the preset archival material at at least one future moment based on the optical flow information; A Kalman filter module is used to obtain the archival motion state of the area in the surveillance image where the material information is the preset archival material, and obtain the predicted archival trajectory of the area where the material information is the preset archival material at at least one future moment based on the archival motion state and the predicted motion trend.

7. The method for monitoring and managing confidential archives compact racks according to claim 2, characterized in that: The material recognition model further includes an environment fusion module, the output of which is connected to the input of the feature extraction module; the environment fusion module is used to fuse the environment parameters pre-obtained based on the image data with the image data.

8. A monitoring and management device for confidential archives compact shelving, characterized in that: include: A processing unit, configured to process the surveillance image obtained by photographing the confidential archives compact rack to obtain image data; an identification unit, configured to input the image data into a material identification model to obtain material information, wherein the material identification model is obtained by marking the material of the historical image data of the confidential archives compact shelving and training based on the marking results and the historical image data; a blurring unit, configured to use a dynamic Gaussian blurring algorithm to perform blurring processing on the area of ​​the surveillance image where the material information is a preset archival material; A prediction unit is configured to utilize the surveillance image and a motion prediction model to obtain a predicted archive trajectory for at least one future moment, and to perform blurring on an area in the surveillance image for at least one future moment in which the material information is a preset archive material according to the predicted archive trajectory, wherein the motion prediction model is trained based on historical archive motion information.

9. An electronic device, characterized in that: including memory and processor; The memory is used to store computer programs; The processor is used to implement the confidential archive compact shelving monitoring and management method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and when the computer program is executed by the processor, the method for monitoring and managing the confidential archive compact shelving according to any one of claims 1 to 7 is implemented.

Images