Intelligent judicial security management and control system

Abstract

The invention discloses an intelligent judicial security and protection management and control system, and relates to the technical field of security and protection management and control. A security and protection monitoring module is arranged to collect monitoring videos of a judicial place in real time, and an abnormality recognition module outputs recognition result data in combination with an abnormal behavior recognition model; the early warning management and control module obtains locking area information of a plurality of behavior periods based on the identification result data of which the behavior signal is 1, and the information analysis module analyzes the frequency of different security areas in the locking area; the maintenance cycle optimization module updates the maintenance levels of all the monitoring devices in all the security and protection areas in combination with the preset maintenance duration of the monitoring devices arranged in each security and protection area, so that on one hand, the maintenance work is more targeted, and unnecessary resource waste is avoided; and on the other hand, the maintenance level is updated in combination with the preset maintenance duration, so that the monitoring equipment in high-frequency use and key security and protection areas can be maintained more timely and more comprehensively.

Description

Technical Field

[0001] The present invention relates to the technical field of security control and management, and in particular to an intelligent judicial security control and management system. Background Art

[0002] In modern society, the safe and stable operation of judicial places is of vital importance. Judicial places usually have several security areas. Due to different security focuses, the specifications of monitoring equipment deployed in each area are also different. For example, in the cell area of ​​the prison, high-definition and all-round monitoring equipment is required to monitor the behavior of prisoners in real time; while in the jury area of ​​the court, only monitoring equipment with general clarity may be required. This method of deploying monitoring equipment of different specifications according to different security needs can effectively improve the security level of judicial places; At present, in order to save maintenance costs, a unified time is usually set to maintain all deployed monitoring equipment. However, monitoring equipment of different specifications may have different maintenance cycles, and monitoring equipment deployed in specific security areas may require more frequent calibration and cleaning, and unified maintenance may not be able to meet these needs in a timely manner. This may result in the inability to obtain effective monitoring information in a timely manner in the event of a security incident, posing potential risks to the safety of judicial venues. In addition, unified maintenance may cause maintenance personnel to concentrate their work pressure, affecting maintenance quality and efficiency. Maintenance personnel may not be able to conduct comprehensive and detailed inspections and maintenance of all equipment within a limited time, leaving potential safety hazards; In order to solve the above problems, the present invention proposes a solution. Summary of the invention

[0003] The purpose of the present invention is to provide an intelligent judicial security control system to solve the problems raised in the above background technology; The purpose of the present invention can be achieved by the following technical solutions: An intelligent judicial security control system, comprising: The abnormality recognition module is used to receive real-time judicial venue surveillance videos and periodically input all stored judicial venue surveillance videos into the abnormal behavior recognition model to obtain recognition result data corresponding to the behavior recognition period, wherein the recognition result data includes a behavior signal, which is one of the numbers 1 or 0; When the behavior signal in the recognition result data of a behavior recognition cycle is 1, the recognition result data also includes the security area where the abnormal behavior is identified and the personnel characteristics of the target personnel whose abnormal behavior is identified; The early warning control module receives the recognition result data with the behavior signal of 1 and generates the locking area information of the behavior recognition cycle corresponding to the recognition reception data and transmits it to the information analysis module; An information analysis module, used to store the locked area information of all received behavior recognition cycles; The information analysis module is also used to analyze the locked area information of all behavior recognition cycles stored in the information analysis cycle based on the preset information analysis rules at each information analysis cycle to obtain the optimized maintenance benchmark information of the information analysis cycle; The maintenance cycle optimization module is used to update the maintenance level of the monitoring equipment deployed in all security areas in the judicial venue based on the preset optimization update rules after receiving the optimized maintenance benchmark information of the information analysis cycle. The maintenance level is divided into three types: primary maintenance, intermediate maintenance and advanced maintenance.

[0004] Further, when the early warning control module receives a recognition result data of a behavior recognition cycle containing a behavior signal of 1, it first starts timing, then obtains the security area containing the recognized abnormal behavior, and retrieves the real-time monitoring video of the security area; Then, according to the security area, monitoring videos of several security areas associated with the security area are synchronously retrieved; Then, the personal features of the target person are extracted from the recognition result data, and the retrieved real-time surveillance videos of all security areas are synchronously displayed to the security control personnel of the judicial venue for viewing. At the same time, based on the extracted personal features of the target person, the target person is queried and locked in real time in the retrieved real-time surveillance videos of all security areas; When the timing reaches the preset warning response time, all security areas locked to the target person within the warning response time are extracted, and the locking area information of the behavior recognition cycle is generated based on them and the security areas where abnormal behavior is identified.

[0005] Furthermore, each time a target person is found in a real-time surveillance video of a retrieved security area, the personal features of the target person are re-identified, and if a feature that does not exist in the personal features of the current target person is identified, the feature is updated into the personal features of the current target person; And according to the updated personal characteristics of the current target person, the target person is searched and locked in the remaining security monitoring videos.

[0006] Furthermore, after receiving the optimized maintenance benchmark information of an information analysis cycle, the optimized update rules for updating the maintenance level of the monitoring equipment deployed in all security areas in the judicial venue are as follows: S21: Obtain all multi - frequency regions included in the optimized maintenance benchmark information for this information analysis period and their maintenance update benchmark quantities for this information analysis period, extract all multi - frequency regions with a maintenance update benchmark quantity greater than or equal to P1 from them, label these extracted multi - frequency regions as K1, K2... Ke in descending order of their maintenance update benchmark quantities, and simultaneously label their corresponding maintenance update benchmark quantities as L1, L2... Le; S22: Use the formula to calculate and obtain the maintenance level evaluation and screening quantity N1 for multi - frequency regions K1, K2... Ke. In the formula, M1, M2... Me are the preset maintenance durations of the monitoring devices deployed in multi - frequency regions K1, K2... Ke in sequence; Compare the magnitudes of N1 and P2. If N1 ≥ P2, then use the formula to calculate and obtain the elimination assessment quantities T1, T2... Te for multi - frequency regions Ke, Ke - 1... K1 in sequence. Here, Lt refers to each of the maintenance update benchmark quantities L1, L2... Le, and Mt refers to each of the preset maintenance durations M1, M2... Me, where P2 is the preset high - level maintenance screening quantity; S23: Select the multi - frequency region corresponding to the largest elimination assessment quantity from the elimination assessment quantities T1, T2... Te, remove this multi - frequency region from multi - frequency regions K1, K2... Ke, calculate the maintenance level evaluation and screening quantity for all the remaining multi - frequency regions after removal, and update and label it as N1. At this time, compare the magnitudes of N1 and P2 again until N1 < P2. Obtain all the multi - frequency regions participating in the calculation of N1 at this time, and update the maintenance levels of these multi - frequency regions to high - level maintenance; S24: Update the maintenance levels of all the multi - frequency regions removed from multi - frequency regions K1, K2... Ke in S23 to medium - level maintenance; S25: Label the remaining multi - frequency regions in the optimized maintenance benchmark information for this information analysis period that have not had their maintenance levels updated as Q1, Q2... Qq in descending order of their corresponding maintenance update benchmark quantities, where q ≥ 1; label the maintenance update benchmark quantities corresponding to multi - frequency regions Q1, Q2... Qq as R1, R2... Rq in sequence; S26: Use the formula to calculate and obtain the maintenance level evaluation and screening quantity U1 for multi - frequency regions Q1, Q2... Qq, where V1, V2... Vq are the preset maintenance durations of the monitoring devices deployed in multi - frequency regions Q1, Q2... Qq in sequence; Compare the magnitudes of U1 and P3. If U1 ≥ P3, then calculate and obtain the rejection evaluation quantities for the multi-frequency regions Q1, Q2... Qq in sequence according to S22, and remove several multi-frequency regions from the multi-frequency regions Q1, Q2... Qq according to step S23 until finally U1 < P3. Obtain all the multi-frequency regions participating in the calculation of U1 at this time, and update the maintenance levels of these multi-frequency regions to intermediate maintenance; S27: Update the maintenance levels of all the remaining multi-frequency regions in the optimized maintenance reference information of this information analysis period, for which the maintenance levels have not been updated, to primary maintenance.

[0007] Advantages of the present invention: The present invention sets up a security monitoring module to collect the monitoring videos of all security regions in the judicial site in real time. The abnormal recognition module combines an abnormal behavior recognition model to recognize the abnormal behaviors of the personnel in the monitoring videos and outputs the recognition result data. The early warning and control module extracts the personnel characteristics of the target personnel with abnormal behaviors identified based on the recognition result data with a behavior signal of 1, and queries and locks the target personnel in several security regions associated with the security regions where the abnormal behaviors are recognized. For each target personnel queried and locked, the personnel characteristics are extracted once based on the corresponding security region where it is queried and locked. In this way, by using the monitoring devices of different specifications in the security regions to continuously improve the personnel characteristics of the target personnel, not only the accuracy and speed of querying and locking are accelerated, but also it is convenient for later search; The present invention combines the security regions where the target personnel are queried and locked through the early warning and control module to obtain the locked region information of several behavior cycles, and the information analysis module analyzes and compares the frequencies of different security regions in the locked regions to obtain the maintenance update reference quantities of all security regions. The maintenance cycle optimization module updates the maintenance levels of all monitoring devices in all security regions in combination with the preset maintenance durations of the monitoring devices arranged in each security region. In this way, the maintenance update reference quantities are determined according to the actual use and recognition situations of different security regions, making the maintenance work more targeted and avoiding unnecessary resource waste. On the other hand, the maintenance levels are updated in combination with the preset maintenance durations to ensure that the monitoring devices in the high-frequency use and key security regions can be maintained more timely and comprehensively; after the update, the new maintenance levels of all monitoring devices are sent to the electronic devices of the maintenance personnel for the maintenance personnel to plan the maintenance. Description of the Drawings

[0008] The present invention will be further described below with reference to the drawings.

[0009] Figure 1 is the system block diagram of the present invention; Figure 2 is the method flow chart of the present invention. Detailed Embodiments

[0010] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0011] like Figure 1 , 2 As shown, an intelligent judicial security control system includes a security monitoring module, an abnormality identification module, an early warning control module, an information analysis module and a maintenance cycle optimization module; The security monitoring module is used to collect surveillance videos of several security areas in the judicial venue in real time. In this embodiment, several security areas are demarcated in the judicial venue based on the functional layout and security requirements of the judicial venue; Each security area is equipped with a number of monitoring devices of different specifications based on its protection focus and security level. In this embodiment, the monitoring device refers to a monitoring camera. The specifications of the monitoring cameras selected in different security areas are different according to the protection focus and security level of the deployment area. The security monitoring module generates a real-time judicial venue monitoring video based on the real-time collected monitoring videos of all security areas in the judicial venue and transmits it to the abnormality recognition module; An abnormality recognition module is used to identify abnormal behavior in surveillance videos of judicial places. The abnormality recognition module pre-stores a trained abnormal behavior recognition model, and the abnormal behavior recognition model can identify abnormal behavior of people in the video; The abnormal behavior recognition model stores the judicial venue surveillance video received in real time and inputs all the judicial venue surveillance videos stored in the behavior recognition cycle into the abnormal behavior recognition model at intervals of a preset behavior recognition cycle, and the abnormal behavior recognition model outputs recognition result data corresponding to the behavior recognition cycle, wherein the recognition result data at least includes a behavior signal. In this embodiment, the behavior signal can only be selected from the numbers 1 or 0, and the behavior signal of 1 indicates that the behavior is recognized to be abnormal, and the behavior signal of 0 indicates that the behavior is not recognized to be abnormal; When the behavior signal is 1, the recognition result data also includes the security area where the abnormal behavior is identified and the personnel characteristics of the target person who has the abnormal behavior, wherein the personnel characteristics include one or more of appearance characteristics, clothing characteristics, behavior characteristics, and carried item characteristics; Appearance features may be one or more of facial features, hairstyle features, and body shape features; clothing features may be one or more of clothing style and accessories features; and behavior features may be one or more of walking posture features and body swing amplitude features; Facial features can be the outline of the facial features: including the shape of the eyes (such as single eyelids, double eyelids, big eyes, small eyes, etc.), the shape of the nose (such as high nose bridge, flat nose bridge, hooked nose, etc.), the shape of the mouth (such as thick lips, thin lips, cherry mouth, etc.), facial proportions: such as the length and width ratio of the face, forehead height, chin shape, etc. Skin color: it can be judged whether the person's skin color is fair, dark, yellow, etc., facial hair: whether there is a beard, the shape and length of the beard, whether there are eyebrows and the thickness of the eyebrows, and special facial marks: such as one or more of moles, birthmarks, scars, etc.; Hair style features can be hair length: long hair, short hair, medium-long hair, etc.; hair style: straight hair, curly hair, wavy hair, etc.; whether there is a ponytail, braid, bun, etc.; hair color: one or more of natural hair colors such as black, brown, blonde, red, or other special dyed colors; Body shape characteristics can be height, weight, and thinness: judging the height of a person by comparing it with surrounding objects, and whether the body shape is fat, strong, thin, etc. Body proportions: such as the ratio of upper and lower body, the ratio of shoulder width to hip width, etc.; Clothing styles include top types: such as T-shirts, shirts, jackets, windbreakers, etc.; bottom types: such as trousers, shorts, skirts, etc.; clothing color: the main color of the overall clothing and the matching color combination; clothing pattern: whether there are one or more patterns such as stripes, plaids, prints, etc.; Accessory features include hats: different styles of hats such as baseball caps, berets, and top hats; glasses: different types of glasses such as sunglasses and myopia glasses, as well as the shape and color of the frames; jewelry: necklaces, bracelets, earrings, rings, etc., including the material and style of the jewelry; backpacks or handbags: one or more of the following features: style, color, size, etc. When there is a behavior signal of 1 in the recognition result data of a behavior recognition cycle, the recognition result data is transmitted to the early warning control module; The early warning control module is used to warn of abnormal behavior, and the early warning control module pre-stores the associated area information of all security areas in the judicial venue; The associated area information of one of the security areas includes several security areas associated with the security area in the judicial venue. When abnormal behavior is identified in the security area, surveillance videos of several security areas associated with the security area are synchronously retrieved; After receiving the identification result data of a behavior identification cycle, the early warning control module first starts timing, then obtains the security area where the abnormal behavior is identified, and retrieves the real-time monitoring video of the security area; Then, according to the security area, monitoring videos of several security areas associated with the security area are synchronously retrieved; Then, the personal features of the target person are extracted from the recognition result data, and the real-time surveillance videos of all the security areas retrieved are synchronously displayed to the security control personnel of the judicial venue for viewing. At the same time, based on the extracted personal features of the target person, the real-time surveillance videos of all the security areas retrieved are searched and locked in real time. When the target person is locked, it is marked with a red frame in the surveillance video; Each time a target person is found in a real-time surveillance video of a retrieved security area, the personal features of the target person are re-identified. If a feature that does not exist in the personal features of the current target person is identified, the feature is updated into the personal features of the current target person; And according to the updated personal characteristics of the current target person, the target person is searched and locked in the remaining security monitoring videos. This method improves the recognition accuracy and recognition rate of the target person, and provides convenience for later viewing; When the timing reaches the preset warning response time, all security areas locked to the target person within the warning response time are extracted, and the locking area information of the behavior recognition cycle is generated based on the security areas where the abnormal behavior is identified; The early warning control module transmits the locked area information of the behavior recognition cycle to the information analysis module for storage; The information analysis module is used to periodically analyze the locked area information of all behavior recognition cycles stored in it. The analysis steps are as follows: S11: Obtain the locked area information of all behavior recognition cycles stored in the information analysis module in one information analysis cycle, extract all security areas with recognized abnormal behaviors contained therein and deduplicate them, and mark all remaining security areas after deduplication as A1, A2...Aa, a≥1; S12: from the locked area information of all behavior recognition cycles obtained in S11, all locked area information in which the security area in which the behavior abnormality is recognized is the security area A1 is obtained and marked as B1, B2...Bb, b≥1; S13: extract all security areas except the security area A1 included in the locked area information B1, B2...Bb and remove duplicates, and mark all the deduplicated security areas as C1, C2...Cc, where c≥1; S14: Obtain the total number D1 of the locked area information including the security area C1 from the locked area information B1, B2...Bb, and compare D1 with D. If D1≥D, the security area C1 is used as the multi-frequency area of ​​the security area A1, otherwise, no processing is performed, wherein D is a preset multi-frequency area screening comparison threshold; S15: According to S14, the total number D2, D3 ... Dc of the locked area information including the security areas C2, C3 ... Cc is obtained from the locked area information B1, B2 ... Bb in sequence, and each time the total number is obtained, the total number is compared with D, and a plurality of multi-frequency areas of the security area A1 are obtained based on the comparison result; S16: according to S13 to S15, a plurality of multi-frequency areas of the security areas A2, A3 ... Aa are obtained in sequence, and the plurality of multi-frequency areas of the security areas A1, A2 ... Aa are deduplicated and all the remaining multi-frequency areas after deduplication are marked as E1, E2 ... Ee, e ≥ 1; S17: Calculate and obtain the maintenance update reference amount J1 of the multi-frequency area E1 in the information analysis cycle according to a preset calculation rule. The calculation rule is as follows: S171: Traverse all multi-frequency areas of security areas A1, A2...Aa in turn, obtain all security areas including the multi-frequency area E1, re-calibrate them as belonging security areas based on the multi-frequency area E1, and mark them as F1, F2...Ff, f≥1 respectively; S172: Obtain the total number of the security areas whose behaviors are identified as abnormal as belonging security areas F1 and including the locked area information of the multi-frequency area E1 from the locked area information of all behavior recognition cycles obtained in S11, and mark them as the frequency statistics G1 of the multi-frequency area E1 based on the belonging security area F1; S173: According to S171 to S172, frequency statistics G2, G3 ... Gf of the multi-frequency area E1 based on the belonging security areas F2, F3 ... Ff are calculated in sequence; S174: Using formula Calculate and obtain the maintenance update benchmark J1 of the multi-frequency area E1 in the information analysis cycle, where I1 is the sum of all multi-frequency areas of the acquired security areas A1, A2...Aa, where H1 is the total number of locked area information of all behavior recognition cycles acquired in S11, and ɑ1 and ɑ2 are the preset first and second dimensional feature proportion factors; S18: according to S17, sequentially calculate and obtain the maintenance and update reference quantities J2, J3, ...Je of the multi-frequency areas E2, E3, ...Ee in the information analysis cycle; The information analysis module generates optimized maintenance benchmark information of the information analysis cycle according to the maintenance optimization benchmark quantities of the multi-frequency areas E1, E2...Ee in the information analysis cycle, and transmits it to the maintenance cycle optimization module; A maintenance cycle optimization module is used to periodically optimize and adjust the monitoring cycle of monitoring equipment in all security areas within the judicial area. The maintenance cycle optimization module stores the maintenance levels of monitoring equipment deployed in all security areas within the current judicial area. In this embodiment, the maintenance levels include three types: primary maintenance, intermediate maintenance, and advanced maintenance. Each type of maintenance level corresponds to a preset maintenance interval. Primary maintenance corresponds to the longest maintenance interval, and advanced maintenance corresponds to the shortest maintenance interval. After receiving the optimized maintenance benchmark information of the information analysis cycle transmitted, the maintenance cycle optimization module optimizes and adjusts the maintenance cycle of the monitoring equipment deployed in all security areas according to the preset optimization rules. The optimization rules are as follows: S21: Obtain all multi-frequency areas included in the optimized maintenance benchmark information of the information analysis period and their maintenance update benchmark amounts in the information analysis period, and extract all multi-frequency areas whose maintenance update benchmark amounts are greater than or equal to P1, and mark these extracted multi-frequency areas as K1, K2...Ke in descending order of their maintenance update benchmark amounts, and mark their corresponding maintenance update benchmark amounts as L1, L2...Le, where P1 is a preset benchmark initial screening amount; S22: Utilize the formula Calculate and obtain the maintenance level assessment screening quantity N1 of the multi-frequency areas K1, K2...Ke, where M1, M2...Me are respectively the preset maintenance durations of the monitoring devices deployed in the multi-frequency areas K1, K2...Ke, and the preset maintenance period of one monitoring device is set by the deployment personnel based on the installation environment and its hardware configuration when the monitoring device is first deployed in the corresponding security area; Compare N1 and P2. If N1 ≥ P2, use the formula Calculate and obtain the rejection assessment quantities T1, T2...Te of the multi-frequency areas Ke, Ke-1...K1 in sequence, where Lt refers to each of the maintenance update benchmark quantities L1, L2...Le, Mt refers to each of the preset maintenance durations M1, M2...Me, and P2 is the preset advanced maintenance screening quantity; S23: Select the multi - frequency region corresponding to the largest rejection evaluation quantity from the rejection evaluation quantities T1, T2... Te, remove this multi - frequency region from the multi - frequency regions K1, K2... Ke, calculate the maintenance level evaluation and screening quantity of all the remaining multi - frequency regions after removal, and update and mark it as N1. At this time, compare the sizes of N1 and P2 again until N1 < P2. Obtain all the multi - frequency regions participating in the calculation of N1 at this time, and update the maintenance levels of these multi - frequency regions to high - level maintenance; S24: Update the maintenance levels of all the multi - frequency regions removed from the multi - frequency regions K1, K2... Ke in S23 to medium - level maintenance; S25: Mark the remaining multi - frequency regions in the optimized maintenance benchmark information of this information analysis period, which have not had their maintenance levels updated, as Q1, Q2... Qq in descending order according to their corresponding maintenance update benchmark quantities, where q≥1; Mark the maintenance update benchmark quantities corresponding to the multi - frequency regions Q1, Q2... Qq as R1, R2... Rq in sequence; S26: Use the formula to calculate and obtain the maintenance level evaluation and screening quantity U1 of the multi - frequency regions Q1, Q2... Qq, where V1, V2... Vq are the preset maintenance durations of the monitoring devices arranged in the multi - frequency regions Q1, Q2... Qq in sequence; Compare the sizes of U1 and P3. If U1≥P3, calculate and obtain the rejection evaluation quantities of the multi - frequency regions Q1, Q2... Qq in sequence according to S22, and remove several multi - frequency regions from the multi - frequency regions Q1, Q2... Qq according to step S23 until the final U1 < P3. Obtain all the multi - frequency regions participating in the calculation of U1 at this time, and update the maintenance levels of these multi - frequency regions to medium - level maintenance, where P3 is the preset medium - level maintenance screening quantity; S27: Update the maintenance levels of the remaining multi - frequency regions in the optimized maintenance benchmark information of this information analysis period, which have not had their maintenance levels updated, to primary maintenance; After the update is completed, send the new maintenance levels of the monitoring devices arranged in all the security regions in the current judicial region to the electronic devices of the maintenance personnel for the maintenance personnel to conduct regular maintenance; In the description of the specification, the descriptions referring to terms such as "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above - mentioned terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0012] The above contents are merely examples and explanations of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the invention or exceed the scope defined by the claims, they shall all fall within the protection scope of the present invention.

[0013] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. An intelligent judicial security control system, characterized in that: include: The security monitoring module is used to collect the surveillance videos of several security areas in the judicial venue in real time, generate real-time judicial venue surveillance videos based on the surveillance videos of all security areas in the judicial venue collected in real time, and transmit them to the abnormality recognition module; The security area is defined within the judicial venue based on the functional layout and security requirements of the judicial venue; Each security area is equipped with several monitoring devices of different specifications based on its protection focus and security level. Monitoring devices refer to monitoring cameras. The specifications of the monitoring cameras selected in different security areas are different according to the protection focus and security level of the deployment area. An abnormality recognition module is used to identify abnormal behavior in surveillance videos of judicial places. The abnormality recognition module pre-stores a trained abnormal behavior recognition model, and the abnormal behavior recognition model can identify abnormal behavior of people in the video; The abnormal behavior recognition model stores the judicial venue surveillance video received in real time and inputs all the judicial venue surveillance videos stored in the behavior recognition cycle into the abnormal behavior recognition model at each preset behavior recognition cycle, and the abnormal behavior recognition model outputs the recognition result data corresponding to the behavior recognition cycle, wherein the recognition result data at least includes a behavior signal, and the behavior signal is 1 or 0, and the behavior signal is 1, indicating that the behavior is abnormal, and the behavior signal is 0, indicating that the behavior is not abnormal; When the behavior signal is 1, the recognition result data also includes the security area where the abnormal behavior is identified and the personnel characteristics of the target person who has the abnormal behavior, wherein the personnel characteristics include one or more of appearance characteristics, clothing characteristics, behavior characteristics, and carried item characteristics; The early warning control module receives the recognition result data with the behavior signal of 1, generates the lock area information of the corresponding behavior recognition cycle according to the recognition result data, and transmits the lock area information to the information analysis module; When the early warning control module receives a recognition result data of a behavior recognition cycle containing a behavior signal of 1, it first starts timing, then obtains the security area where the abnormal behavior is identified contained in the recognition result data, and retrieves the real-time monitoring video of the security area; Then, according to the security area, monitoring videos of several security areas associated with the security area are synchronously retrieved; Then, the personal features of the target person are extracted from the recognition result data, and the retrieved real-time surveillance videos of all security areas are synchronously displayed to the security control personnel of the judicial venue for viewing. At the same time, based on the extracted personal features of the target person, the target person is queried and locked in real time in the retrieved real-time surveillance videos of all security areas; When the timing reaches the preset warning response time, all security areas locked to the target person within the warning response time are extracted, and the locking area information of the behavior recognition cycle is generated according to all security areas and the security areas where abnormal behavior is identified; Each time a target person is found in a real-time surveillance video of a retrieved security area, the personal features of the target person are re-identified. If a feature that does not exist in the personal features of the current target person is identified, the non-existent feature is updated into the personal features of the current target person; And according to the updated personal characteristics of the current target person, the target person is searched and locked in the remaining security monitoring videos; An information analysis module, used to store the locked area information of all received behavior recognition cycles; The information analysis module is also used to analyze the locked area information of all behavior recognition cycles stored in the information analysis cycle based on the preset information analysis rules at each information analysis cycle to obtain the optimized maintenance benchmark information of the information analysis cycle; The information analysis rules for obtaining the optimized maintenance benchmark information for an information analysis cycle are as follows: S11: Obtain the locked area information of all behavior recognition cycles stored in the information analysis module in one information analysis cycle, extract all security areas with recognized abnormal behaviors contained therein and perform deduplication, and mark all remaining security areas after deduplication as A1, A2...Aa, a≥1; S12: Obtain all the locked area information of the security area A1 where the security area identified as abnormal behavior is the security area from the locked area information of all behavior recognition cycles obtained in S11 and mark them as B1, B2...Bb, b≥1; S13: extract all security areas except the security area A1 included in the locked area information B1, B2...Bb and remove duplicates, and mark all the deduplicated security areas as C1, C2...Cc, c≥1; S14: Obtain the total number D1 of the locked area information including the security area C1 from the locked area information B1, B2...Bb, and compare D1 with D. If D1≥D, the security area C1 is used as the multi-frequency area of ​​the security area A1, otherwise, no processing is performed, wherein D is a preset multi-frequency area screening comparison threshold; S15: According to S14, the total number D2, D3 ... Dc of the locked area information including the security areas C2, C3 ... Cc is obtained from the locked area information B1, B2 ... Bb in sequence, and each time the total number is obtained, the total number is compared with D, and a plurality of multi-frequency areas of the security area A1 are obtained based on the comparison result; S16: according to S13 to S15, a plurality of multi-frequency areas of the security areas A2, A3 ... Aa are obtained in sequence, and the plurality of multi-frequency areas of the security areas A1, A2 ... Aa are deduplicated and all the remaining multi-frequency areas after deduplication are marked as E1, E2 ... Ee, e ≥ 1; S17: Calculate and obtain the maintenance update reference amount J1 of the multi-frequency area E1 in the information analysis cycle according to a preset calculation rule. The calculation rule is as follows: S171: Traverse all multi - frequency regions of security regions A1, A2... Aa in sequence, obtain all security regions that contain multi - frequency region E1, and re - label these security regions as belonging security regions based on multi - frequency region E1, and mark them as F1, F2... Ff respectively, where f≥1; S172: Obtain the security region with abnormal behavior recognition as the belonging security region F1 from the locked region information of all behavior recognition cycles obtained in S11, and the total number of locked region information containing multi - frequency region E1, and label this total number as the frequency statistic G1 of multi - frequency region E1 based on the belonging security region F1; S173: Calculate and obtain the frequency statistics G2, G3... Gf of multi - frequency region E1 based on the belonging security regions F2, F3... Ff in sequence according to S171 to S172; S174: Using formula Calculate and obtain the maintenance update benchmark J1 of the multi-frequency area E1 in the information analysis cycle, where I1 is the sum of all multi-frequency areas of the acquired security areas A1, A2...Aa, H1 is the total number of locked area information of all behavior recognition cycles acquired in S11, ɑ1 and ɑ2 are the preset first and second dimensional feature proportion factors; S18: Calculate and obtain the maintenance and update reference quantities J2, J3... Je of multi - frequency regions E2, E3... Ee in this information analysis cycle in sequence according to S17; The information analysis module generates the optimized maintenance reference information for this information analysis cycle based on the maintenance and optimization reference quantities of multi - frequency regions E1, E2... Ee in this information analysis cycle; The maintenance cycle optimization module is used to update the maintenance level of the monitoring devices deployed in all security regions within the judicial venue based on a preset optimization and update rule after receiving the optimized maintenance reference information for the information analysis cycle. The maintenance levels are divided into three types: primary maintenance, intermediate maintenance, and advanced maintenance; The optimization and update rule for updating the maintenance level of the monitoring devices deployed in all security regions within the judicial venue after receiving the optimized maintenance reference information for an information analysis cycle is as follows: S21: Obtain all multi - frequency regions included in the optimized maintenance reference information for this information analysis cycle and the maintenance and update reference quantities of these all multi - frequency regions in this information analysis cycle, and extract all multi - frequency regions whose maintenance and update reference quantities are greater than or equal to P1. Mark these extracted multi - frequency regions as K1, K2... Ke in sequence according to the descending order of the maintenance and update reference quantities, and mark the corresponding maintenance and update reference quantities of these multi - frequency regions as L1, L2... Le respectively; S22: Utilize the formula Calculate and obtain the maintenance level assessment screening quantity N1 of the multi-frequency areas K1, K2...Ke, where M1, M2...Me are respectively the preset maintenance durations of the monitoring equipment deployed in the multi-frequency areas K1, K2...Ke; Compare N1 and P2. If N1 ≥ P2, use the formula Calculate and obtain the rejection assessment quantities T1, T2...Te of the multi-frequency areas Ke, Ke-1...K1 in sequence, where Lt refers to each of the maintenance update benchmark quantities L1, L2...Le, Mt refers to each of the preset maintenance durations M1, M2...Me, and P2 is the preset advanced maintenance screening quantity; S23: Select the multi - frequency region corresponding to the largest rejection assessment quantity from the rejection assessment quantities T1, T2... Te, remove this multi - frequency region from the multi - frequency regions K1, K2... Ke, and calculate the maintenance level assessment and screening quantity of all the remaining multi - frequency regions after removal. Update and mark this maintenance level assessment and screening quantity as N1. At this time, compare the sizes of N1 and P2 again until N1 < P2. Obtain all the multi - frequency regions participating in the calculation of N1 at this time, and update the maintenance levels of these multi - frequency regions to advanced maintenance; S24: Update the maintenance levels of all the multi - frequency regions removed from the multi - frequency regions K1, K2... Ke in S23 to intermediate maintenance; S25: Mark all the remaining multi - frequency regions in the optimized maintenance benchmark information of this information analysis period that have not updated the maintenance level as Q1, Q2... Qq in descending order according to their corresponding maintenance update benchmark amounts, where q ≥ 1; mark the maintenance update benchmark amounts corresponding to the multi - frequency regions Q1, Q2... Qq as R1, R2... Rq in sequence. S26: Utilize the formula Calculate and obtain the maintenance level assessment screening quantity U1 of the multi-frequency areas Q1, Q2...Qq, where V1, V2...Vq are respectively the preset maintenance durations of the monitoring equipment deployed in the multi-frequency areas Q1, Q2...Qq; Compare the magnitudes of U1 and P3. If U1 ≥ P3, calculate and obtain the elimination assessment amounts of the multi - frequency regions Q1, Q2... Qq in sequence according to S22, and eliminate several multi - frequency regions from the multi - frequency regions Q1, Q2... Qq until finally U1 < P3. Obtain all the multi - frequency regions participating in the calculation of U1 at this time, and update the maintenance levels of these multi - frequency regions to intermediate maintenance. S27: Update the maintenance levels of all the remaining multi - frequency regions in the optimized maintenance benchmark information of this information analysis period that have not updated the maintenance level to primary maintenance.

2. The intelligent judicial security control system according to claim 1, characterized in that: After the maintenance levels of all the monitoring devices deployed in all the security regions within the judicial venue have been updated once, send the new maintenance levels of all the monitoring devices deployed in all the security regions within the current judicial region to the electronic devices of the maintenance personnel.

Images