Building intelligent security monitoring system

Abstract

The invention discloses an intelligent building security and protection monitoring system, and relates to the technical field of security and protection monitoring. Performing feature data extraction of different degrees on the video image information with the abnormal behavior to obtain a first feature video image and a second feature video image; dividing the second feature video image into a target video image and a locked video image according to the condition of the abnormal behavior in the second feature video image; wherein the target video image is a video image having an abnormal behavior, and the locked video image is a video image suspected to have an abnormal behavior; extracting an attention video image corresponding to the target video image from the first feature video image, and extracting a positioning video image corresponding to the locking video image from the first feature video image; the effect is that the accuracy of abnormal behavior recognition is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of security monitoring, and more particularly to an intelligent building security monitoring system. Background Art

[0002] With the acceleration of urbanization and the continuous expansion of building scale, building security issues are becoming increasingly prominent. Traditional security monitoring systems, which rely primarily on manual monitoring and simple video recording, have gradually exposed many limitations when faced with complex and ever-changing security scenarios, making it difficult to meet the high security standards required by modern buildings.

[0003] With technological advancements, security systems with basic intelligent analysis capabilities have begun to emerge, such as simple rule-based motion detection systems. These systems can detect the movement of objects within an image, but are limited to determining whether an object is moving and are unable to accurately distinguish between normal and abnormal behavior. Furthermore, some existing security monitoring systems lack comprehensive consideration of behavior diffusion trends and environmental factors when analyzing abnormal behavior. In scenarios such as building fires and violent incidents, abnormal behavior often spreads. Traditional systems fail to fully consider the diffusion rate of abnormal behavior and the impact of environmental factors (such as spatial layout and interference factors) on diffusion. This makes it impossible to accurately and proactively warn of the development of abnormal behavior, resulting in the delay in implementing effective prevention and control measures, which may further expand the impact of the incident. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide an intelligent building security monitoring system.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] An intelligent building security monitoring system, comprising:

[0007] Acquisition and extraction module: collects video image information of the building security monitoring area, extracts feature data of different degrees from video image information with abnormal behavior to obtain a first feature video image and a second feature video image;

[0008] A division module: divides the second feature video image into a target video image and a locked video image according to the abnormal behavior in the second feature video image; wherein the target video image is a video image with abnormal behavior, and the locked video image is a video image suspected of having abnormal behavior;

[0009] Image extraction module: extracts the attention video image corresponding to the target video image from the first feature video image, and extracts the positioning video image corresponding to the locked video image from the first feature video image;

[0010] Combination module: combines the abnormal behavior location points determined in the video image of interest with the suspected abnormal location points in the positioning video image to obtain a combined location area;

[0011] Processing and analysis module: processes and analyzes the abnormal behavior location points and suspected abnormal location points in the combined location area to obtain diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location points;

[0012] Judgment module: Process and analyze the diffusion rate ratio data 1 and the diffusion rate ratio data 2 to determine whether the area where the suspected abnormal location point in the combined location area is located is an abnormal behavior location area.

[0013] Preferably, processing and analyzing the abnormal behavior location points and the suspected abnormal location points in the combined location area to obtain diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location points specifically includes the following steps:

[0014] The correlation degree is obtained by extracting the correlation degree between the abnormal behavior location point and the suspected abnormal location point in the combined location area;

[0015] The abnormal behavior diffusion rate within the time period 1 of the abnormal behavior location point determined in the combined location area is statistically analyzed to obtain a first abnormal behavior diffusion rate;

[0016] The rate impact factor of the combined location area is collected, and the diffusion rate ratio data 1 and the diffusion rate ratio data of the suspected abnormal location point are obtained by processing according to the correlation degree, the diffusion rate of the first abnormal behavior and the rate impact factor.

[0017] Preferably, obtaining a first abnormal behavior diffusion rate by statistically analyzing the abnormal behavior diffusion rate within a time period 1 of the abnormal behavior location point determined in the combined location area specifically includes the following steps:

[0018] Compare the correlation degree with a preset correlation threshold, and obtain historical period detection data when the correlation degree is greater than or equal to the preset correlation threshold;

[0019] The historical period detection data and the abnormal behavior diffusion rate of the abnormal location point determined in the combined location area are processed to obtain a first abnormal behavior diffusion rate.

[0020] Preferably, the abnormal behavior diffusion rate of the abnormal location point determined in the historical period detection data and the combined location area is processed to obtain a first abnormal behavior diffusion rate, which specifically includes the following steps:

[0021] Extracting first target detection data within a time period 1 from the historical time period detection data; wherein the first target detection data includes abnormal behavior feature change data 1 and abnormal behavior impact range change data 1;

[0022] The abnormal behavior diffusion rate of the abnormal position point determined in the combined position area is statistically analyzed according to the first target detection data to obtain a first abnormal behavior diffusion rate.

[0023] Preferably, collecting the rate impact factor of the combined location area specifically includes the following steps:

[0024] Calculate the distance between the abnormal location point and the suspected abnormal location point in the combined location area and output the distance data information;

[0025] Detecting interference degree value data information of the abnormal behavior connection area between the determined abnormal location point and the suspected abnormal location point in the combined location area;

[0026] The distance data information and the interference level value data information are combined into a rate impact factor.

[0027] Preferably, the process of obtaining diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location point according to the correlation degree, the diffusion rate of the first abnormal behavior, and the rate influencing factor specifically includes the following steps:

[0028] The abnormal behavior diffusion rate of the suspected abnormal location point in the combined location area is predicted according to the correlation degree, the first abnormal behavior diffusion rate and the rate influencing factor to obtain a second predicted diffusion rate;

[0029] Predicting a third predicted diffusion rate of a suspected abnormal location point in the combined location area based on the first target detection data;

[0030] Processing the historical period detection data to obtain a fourth predicted diffusion rate of the suspected abnormal location point in the combined location area;

[0031] calculating a ratio between the second predicted diffusion rate and the third predicted diffusion rate to obtain diffusion rate ratio data one;

[0032] The ratio between the third predicted diffusion rate and the fourth predicted diffusion rate is calculated to obtain diffusion rate ratio data 2.

[0033] Preferably, processing the historical period detection data to obtain a fourth predicted diffusion rate of the suspected abnormal location point in the combined location area specifically includes the following steps:

[0034] Extracting second target detection data within period 2 from the historical period detection data; wherein the second target detection data includes abnormal behavior feature change data 2 and abnormal behavior impact range change data 2;

[0035] A fourth predicted diffusion rate of the suspected abnormal position point in the combined position area is predicted based on the second target detection data.

[0036] Preferably, the diffusion rate ratio data 1 and the diffusion rate ratio data 2 are processed and analyzed to determine whether the area where the suspected abnormal location point in the combined location area is located is an abnormal behavior location area, specifically comprising the following steps:

[0037] The change ratio between the diffusion rate ratio data 1 and the diffusion rate ratio data 2 is calculated to obtain a change ratio value;

[0038] Preset a diffusion rate change ratio threshold, and compare the change ratio value with the preset diffusion rate change ratio threshold;

[0039] If the change ratio value is greater than the preset diffusion rate change ratio threshold, the suspected abnormal location point in the combined location area is determined to be an abnormal behavior location area;

[0040] If the change ratio value is less than or equal to the preset diffusion rate change ratio threshold, the suspected abnormal position point in the combined position area is determined to be a non-abnormal behavior position area.

[0041] Preferably, the method further comprises the following steps:

[0042] Detect and count the environmental change values ​​of the building security monitoring area during time periods 1, 2, and 3 for suspected abnormal location points in the combined location area based on the non-abnormal behavior location area;

[0043] Preset an environmental change degree warning value, and compare the environmental change degree value with the preset environmental change degree warning value;

[0044] If the environmental change degree value is greater than or equal to the preset environmental change degree warning value, the abnormal behavior risk warning result is output;

[0045] If the environmental change degree value is less than the preset environmental change degree warning value, the abnormal behavior risk warning result is output.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] This method extracts features from video images in a hierarchical manner, starting with the first feature image, which captures basic appearance and motion features for preliminary screening. The second feature image then mines refined semantic features for accurate judgment, leading to a progressively deeper analysis. Simultaneously, based on the abnormal behavior in the second feature image, the method meticulously divides the target video image (clearly abnormal) and the targeted video image (suspected abnormality), effectively reducing misjudgments and missed detections. This significantly improves the accuracy of abnormal behavior recognition, enabling accurate distinctions between different types of abnormal behavior, such as theft and violent conflict.

[0048] This method's analysis of the diffusion rate of suspected anomaly locations can proactively detect trends in abnormal behavior, enabling early warning and prevention before serious consequences arise. Furthermore, for areas identified as non-abnormal, in-depth statistical analysis of environmental change levels is performed. By comparing these values ​​with pre-set warning levels, the system can discern potential anomaly risks arising from environmental changes, even when no obvious abnormal behavior is currently detected.

[0049] When analyzing the spread of abnormal behavior, we fully consider multiple factors, including correlation, diffusion rate, and rate-influencing factors. Correlation measures the closeness of connections between abnormal locations; the first abnormal behavior diffusion rate reflects the development speed of confirmed anomalies; and the rate-influencing factor integrates spatial distance and interference level, comprehensively considering the impact of multiple factors on the spread of anomalies, including space, time, behavioral characteristics, and environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The present invention proposes a module diagram of an intelligent building security monitoring system;

[0051] Figure 2 A schematic diagram of the steps for obtaining a rate influencing factor in a building intelligent security monitoring system proposed by the present invention;

[0052] Figure 3 The present invention provides a schematic diagram of the steps of obtaining diffusion rate ratio data 1 and diffusion rate ratio data 2 in a building intelligent security monitoring system. DETAILED DESCRIPTION

[0053] Reference Figures 1 to 3 .

[0054] The embodiment further illustrates the intelligent building security monitoring system proposed by the present invention.

[0055] An intelligent building security monitoring system, comprising:

[0056] Acquisition and extraction module: collects video image information of the building security monitoring area, extracts feature data of different degrees from video image information with abnormal behavior to obtain a first feature video image and a second feature video image;

[0057] A division module: divides the second feature video image into a target video image and a locked video image according to the abnormal behavior in the second feature video image; wherein the target video image is a video image with abnormal behavior, and the locked video image is a video image suspected of having abnormal behavior;

[0058] Image extraction module: extracts the attention video image corresponding to the target video image from the first feature video image, and extracts the positioning video image corresponding to the locked video image from the first feature video image;

[0059] Combination module: combines the abnormal behavior location points determined in the video image of interest with the suspected abnormal location points in the positioning video image to obtain a combined location area;

[0060] Processing and analysis module: processes and analyzes the abnormal behavior location points and suspected abnormal location points in the combined location area to obtain diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location points;

[0061] Judgment module: Process and analyze the diffusion rate ratio data 1 and the diffusion rate ratio data 2 to determine whether the area where the suspected abnormal location point in the combined location area is located is an abnormal behavior location area.

[0062] The acquisition and extraction module of this application collects video image information of the building security monitoring area, and performs different degrees of feature extraction on video images with abnormal behavior to obtain a first feature video image (including basic features) and a second feature video image (including more detailed features).

[0063] In intelligent building security monitoring scenarios, surveillance cameras collect a large amount of video image information. Video images showing abnormal behavior contain rich details, but not all of this information is useful for identifying anomalies. Feature data extraction can filter out key information from complex images.

[0064] First feature video image: This is usually extracted using relatively basic, broad algorithms. For example, edge detection algorithms (such as the Canny operator) are used to outline the edges of objects in the video image and determine the approximate shape of the object. Optical flow methods are used to track the motion trajectory of objects in the image and obtain basic motion information such as their direction and speed. This includes the basic appearance and motion characteristics of the target. For example, if someone is found running abnormally during monitoring, the first feature video image can show information such as the person's outline, running direction, and approximate speed. These feature extractions are relatively simple and fast, with low computational complexity. They can quickly locate potentially abnormal targets and provide initial clues for subsequent, more in-depth analysis. They serve as the basis for preliminary screening and rapid positioning. When the monitoring system processes a large number of video streams in real time, the first feature video image can be used to quickly identify targets with potential abnormal behavior, narrowing the scope of subsequent analysis and improving system response speed.

[0065] Second-feature video images: Further in-depth analysis based on the first-feature video images. Deep learning techniques such as convolutional neural networks (CNNs) are used to extract detailed features of the target. For example, these techniques analyze a person's body movements (e.g., whether they display actions such as attacking or snatching) and identify facial expressions (e.g., whether they display unusual emotions such as fear or anger). For objects, these techniques can identify attributes such as color and material. These features are more refined, specific, and semantic. For example, in the monitoring of theft, the second-feature video image not only identifies the person's movements but also combines information such as facial expressions and the state of surrounding objects to accurately determine whether the behavior is abnormal. Compared to the first-feature video image, it provides more critical information for determining the nature and severity of abnormal behavior. This allows for precise identification and classification of abnormal behavior. Analysis of the second-feature video image accurately distinguishes different types of abnormal behavior, such as theft, fire, and illegal intrusion, providing a reliable basis for subsequent targeted action.

[0066] The classification module divides the second feature video image into target video images (clearly showing abnormal behavior) and locked video images (suspected of abnormal behavior) based on the specific circumstances of the abnormal behavior, thereby achieving preliminary classification and screening of abnormal behavior videos.

[0067] In the field of building security monitoring, it's necessary to first define standards for various types of abnormal behavior. These standards cover behavioral patterns, movement characteristics, and occurrence scenarios. For example, entering a restricted area is considered an illegal intrusion anomaly; fighting or violent attacks in public areas are considered violent behavior anomalies. The system uses these preset standards combined with detailed features extracted from the second feature video image (such as body movements, facial expressions, and object attributes) to determine whether abnormal behavior exists.

[0068] Identifying Clearly Abnormal Behavior: When the behavioral characteristics in the second feature video image closely match the pre-defined clear abnormal behavior criteria, the video image is classified as a target video image. For example, image analysis technology can detect that the body movements of the characters in the video image meet the pre-defined fighting action standards, such as punching, kicking, etc., and the scene also meets the setting of a public area where violent conflict may occur, and the facial expression also shows relevant emotional characteristics such as anger. In this case, the system can determine that there is a clear violent behavior anomaly, and the corresponding video image will be classified as a target video image.

[0069] The system assesses the confidence level of abnormal behavior identification results. Through multi-dimensional feature fusion analysis, such as combining action continuity, frequency, and surrounding environmental factors, the system calculates the reliability of abnormal behavior judgments. When the confidence level reaches or exceeds a certain threshold (e.g., 90%), abnormal behavior is determined and classified as a target video image. This effectively avoids misjudgments due to a single feature and ensures the accuracy of classification.

[0070] Suspected Abnormal Behavior Determination: When the behavioral characteristics in the second feature video image do not fully meet the criteria for clear abnormal behavior, but exhibit some suspicious signs, it is classified as a locked video image. For example, if a person wanders around a specific area within the monitoring area for an extended period of time, the wandering behavior itself cannot be definitively determined to be abnormal (there may be legitimate reasons), but this behavior does not conform to normal traffic or stay patterns and may be potentially abnormal. When the system detects such ambiguous behavioral characteristics, it classifies the corresponding video image as a locked video image.

[0071] For suspected abnormal behavior, the system further analyzes its uncertainty, taking into account factors such as the time, location, and frequency of the behavior. For example, if someone is loitering around a building late at night, during non-business hours, the likelihood of an abnormality is higher than if they are loitering during normal business hours. Through this comprehensive analysis, the suspected abnormal behavior is identified and the relevant video images are classified as locked video images for subsequent monitoring and in-depth analysis.

[0072] The image extraction module extracts the attention video image corresponding to the target video image and the positioning video image corresponding to the locked video image from the first feature video image, and locates the video content related to different abnormality types.

[0073] The combination module integrates the abnormal behavior location points determined in the focus video image and the suspected abnormal location points in the positioning video image to form a combined location area, which defines the scope for subsequent analysis.

[0074] The processing and analysis module performs processing and analysis on abnormal behavior location points and suspected abnormal location points within the combined location area. Through a series of calculations (such as extracting correlation, statistical abnormal behavior diffusion rate, acquisition rate influencing factors, etc.), it obtains diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location points, and quantifies the diffusion-related characteristics of the abnormal behavior.

[0075] The judgment module further processes and analyzes the diffusion rate ratio data 1 and the diffusion rate ratio data 2, calculates the change ratio of the two and compares it with the preset threshold to determine whether the area where the suspected abnormal location point in the combined location area is located belongs to the abnormal behavior location area, and obtains the final abnormality judgment result.

[0076] Processing and analyzing the abnormal behavior location points and the suspected abnormal location points in the combined location area to obtain diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location points specifically includes the following steps:

[0077] The correlation degree is obtained by extracting the correlation degree between the abnormal behavior location point and the suspected abnormal location point in the combined location area;

[0078] The abnormal behavior diffusion rate within the time period 1 of the abnormal behavior location point determined in the combined location area is statistically analyzed to obtain a first abnormal behavior diffusion rate;

[0079] The rate impact factor of the combined location area is collected, and the diffusion rate ratio data 1 and the diffusion rate ratio data of the suspected abnormal location point are obtained by processing according to the correlation degree, the diffusion rate of the first abnormal behavior and the rate impact factor.

[0080] This application analyzes and determines the similarity or correlation between abnormal behavior location points and suspected abnormal location points in terms of space, time and behavioral characteristics within the combined location area. For example, if two location points are close in space, the time when abnormal behavior occurs is close in time, or the behavior patterns are similar (such as both are fast movements in the same direction), the correlation degree is calculated by quantifying these factors. The degree of correlation can be comprehensively calculated by using spatial distance metrics (such as Euclidean distance) combined with time difference, behavioral feature similarity (using deep learning models to calculate the cosine similarity of feature vectors) and other methods.

[0081] Determining the physical distance between the location of the abnormal behavior and the suspected abnormal location is a key consideration. For example, in building security monitoring, if two points are in adjacent rooms or on similar floors, the closer the distance, the higher the likelihood of a connection. This can be quantified by calculating the Euclidean distance between the two points in a two-dimensional plane (such as the coordinates of the monitoring screen) or in three-dimensional space (incorporating the spatial structure of the building). The closer the distance, the higher the likelihood of a connection.

[0082] In addition to distance, the orientation relationship between location points also plays a role. For example, two location points located one behind the other on the same channel are more likely to have a behavioral correlation than those located on different channels. This positional relationship can be described using parameters such as coordinate azimuth and factored into correlation calculations.

[0083] The time interval between the occurrence of abnormal behavior at two locations is crucial. A short interval, such as abnormal wandering behavior occurring in adjacent areas within a short period of time, indicates a high likelihood of correlation. Based on the timestamps, the difference between the occurrence times of the abnormal behavior at the two locations is calculated. The smaller the difference, the higher the temporal correlation.

[0084] The order in which unusual behaviors occur at two locations also has significance. For example, if an item is lost at one location and then a person is seen hurriedly leaving with a similar item at a nearby location shortly thereafter, this temporal sequence of behaviors suggests a close connection. Analyzing temporal sequences can help determine causal or correlational relationships.

[0085] Determine whether the actions at the abnormal behavior location and the suspected abnormal location are similar. For example, they may both be running fast or have specific hand gestures. Using motion recognition technology, we extract the action feature vectors. We then measure the similarity of the actions by calculating the cosine similarity between the feature vectors. The higher the similarity, the greater the correlation.

[0086] In addition to actions, the consistency of behavioral patterns also needs to be considered. For example, are there periodic repetitive abnormal behaviors (such as periodically wandering around a specific area) or are they triggered by specific environmental factors (such as unusual reactions to specific people). By analyzing and matching behavioral patterns, the degree of correlation can be determined.

[0087] Weighted summation method: Assign weights to each factor in the spatial, temporal, and behavioral feature dimensions (weights can be determined based on the importance of the actual security scenario through expert experience or machine learning training). For example, a weight of 0.4 is assigned to spatial distance, 0.3 to behavioral time interval, and 0.3 to action similarity. Calculate the quantitative value of each factor separately, then perform a weighted sum based on the weights to obtain the correlation value. For example, if the spatial distance quantification value is 0.8 (close distance), the behavioral time interval quantification value is 0.7 (short interval), and the action similarity quantification value is 0.6 (high similarity), then the correlation degree = 0.4 × 0.8 + 0.3 × 0.7 + 0.3 × 0.6 = 0.71.

[0088] Using clustering algorithms (such as DBSCAN) or classification algorithms (such as support vector machines (SVMs), factors such as spatial, temporal, and behavioral characteristics are input into the model as feature vectors. Through model training, the model learns association patterns and outputs association judgment results (such as high association, medium association, and low association categories), or directly outputs an association value between 0 and 1. For example, an SVM model can be trained using a large amount of sample data with known associations, and then new location feature vectors can be input to obtain association predictions.

[0089] The diffusion of abnormal behavior within a specific time period is statistically analyzed for the abnormal behavior location points determined in the combined location area. The diffusion here can be manifested as the expansion of the scope of influence of abnormal behavior (such as the increase in the area of ​​fire spread, the expansion of the area of ​​crowd panic), the change in the intensity of abnormal behavior (such as the escalation of the intensity of violent behavior), etc. By recording the relevant indicator data at different time points in time period one, the first abnormal behavior diffusion rate is obtained using time series analysis methods (such as calculating the ratio of the change in indicators at adjacent time points to the time interval). The first abnormal behavior diffusion rate reflects the development speed of the confirmed abnormal behavior, provides a reference benchmark for judging the diffusion of suspected abnormal location points, and is used to measure the dynamic changes of abnormal behavior in the current area.

[0090] The rate impact factor consists of two parts: first, the distance between the anomalous location and the suspected anomalous location. A closer distance may make the anomaly more likely to spread. Second, the interference level in the region connecting the two detected anomalous behavior. Interference levels include physical obstructions (such as obstacles) and environmental factors (such as strong winds affecting smoke diffusion). Lower interference levels are more conducive to anomaly diffusion. The distance data and interference level data are combined (e.g., through weighted summation) to form the rate impact factor. The rate impact factor comprehensively considers the impact of spatial and environmental factors on anomaly diffusion and is used to adjust and correct the calculation results when subsequently calculating the diffusion rate of suspected anomalous locations.

[0091] Based on the previously obtained correlation degree, the first abnormal behavior diffusion rate, and the rate influencing factor, a specific mathematical model (such as a prediction model based on multiple linear regression or a neural network) is used to predict the abnormal behavior diffusion rate of the suspected abnormal location point in the combined location area to obtain a second predicted diffusion rate. At the same time, based on the first target detection data (including abnormal behavior characteristic change data 1 and abnormal behavior influence range change data 1), the diffusion rate of the suspected abnormal location point is also predicted to obtain a third predicted diffusion rate. The fourth predicted diffusion rate is then obtained by processing the historical time period detection data. The ratio between the second predicted diffusion rate and the third predicted diffusion rate is then calculated to obtain diffusion rate ratio data 1, and the ratio between the third predicted diffusion rate and the fourth predicted diffusion rate is calculated to obtain diffusion rate ratio data 2.

[0092] Diffusion rate ratio data one and two reflect the relationship between the diffusion rate of the suspected abnormal location and other related diffusion rates from different angles. By comparing these ratio data, the diffusion trend and relative change of the suspected abnormal location can be analyzed.

[0093] The method of obtaining a first abnormal behavior diffusion rate by statistically analyzing the abnormal behavior diffusion rate within a time period 1 of the abnormal behavior location point determined in the combined location area specifically includes the following steps:

[0094] Compare the correlation degree with a preset correlation threshold, and obtain historical period detection data when the correlation degree is greater than or equal to the preset correlation threshold;

[0095] The historical period detection data and the abnormal behavior diffusion rate of the abnormal location point determined in the combined location area are processed to obtain a first abnormal behavior diffusion rate.

[0096] The correlation reflects the degree of correlation between the confirmed abnormal behavior location and the suspected abnormal location within the combined location area, in terms of space, time, and behavioral characteristics. The preset correlation threshold is a standard value set based on the data. Comparing the calculated correlation with the correlation threshold is used to screen for cases with a high likelihood of correlation. For example, if the correlation threshold is set to 0.6, a calculated correlation greater than or equal to 0.6 indicates a strong correlation between the two locations, requiring further analysis. If it is less than 0.6, the correlation may be considered weak and subsequent complex analysis may not be performed.

[0097] This comparison can filter out cases with low correlation, avoid unnecessary processing of large amounts of irrelevant or weakly correlated data, improve analysis efficiency, and concentrate resources on in-depth research on highly correlated abnormal behaviors.

[0098] When the correlation is greater than or equal to the preset correlation threshold, it indicates that there may be a close connection between the confirmed abnormal behavior location and the suspected abnormal behavior location. Historical period detection data refers to abnormal behavior data recorded in the building security monitoring area or similar scenarios during similar time periods in the past, including information such as the time, location, behavior pattern, and diffusion of the abnormal behavior. This data can be retrieved and extracted from the system's historical database according to specific criteria (such as the time range and location of the current abnormal behavior).

[0099] Historical detection data provides a comparison and reference for analyzing the current rate of spread of abnormal behavior. By comparing it with historical data, we can understand whether the current trend of abnormal behavior conforms to past patterns and determine whether the spread rate is experiencing normal fluctuations or abnormal changes. This helps to more accurately assess the severity and development trend of abnormal behavior.

[0100] After acquiring historical detection data, comprehensive processing is performed on the current abnormal behavior diffusion situation at the identified abnormal location within the combined location area. Statistical analysis methods can be used, such as calculating the mean, median, and other statistical quantities of the abnormal behavior diffusion rate in the historical data and comparing them with the diffusion rate at the currently identified abnormal location. Alternatively, time series analysis methods can be used, treating historical and current data as a time series and building models (such as ARIMA models) to predict and analyze the current abnormal behavior diffusion rate. Machine learning regression algorithms (such as linear regression) can also be used, using various historical data features (such as time, location, and behavior type) as independent variables and the diffusion rate as the dependent variable to train a model. The trained model can then be used to predict the current first abnormal behavior diffusion rate.

[0101] By processing historical detection data and current data, we can obtain a relatively accurate and comprehensive value that reflects the abnormal behavior diffusion rate of the currently determined abnormal behavior location within period one, namely the first abnormal behavior diffusion rate. This value provides key basic data for further analysis of the diffusion of suspected abnormal locations and for determining the abnormal behavior trend of the entire combined location area.

[0102] The abnormal behavior diffusion rate of the abnormal location point determined in the historical period detection data and the combined location area is processed to obtain a first abnormal behavior diffusion rate, specifically comprising the following steps:

[0103] Extracting first target detection data within a time period 1 from the historical time period detection data; wherein the first target detection data includes abnormal behavior feature change data 1 and abnormal behavior impact range change data 1;

[0104] The abnormal behavior diffusion rate of the abnormal position point determined in the combined position area is statistically analyzed according to the first target detection data to obtain a first abnormal behavior diffusion rate.

[0105] The historical period detection data contains a large amount of abnormal behavior related information at different times in the past. In order to more accurately analyze the abnormal behavior diffusion rate of the abnormal location point in the current combined location area, it is necessary to filter out the data related to the current period (period one), that is, the first target detection data. Relevant records are retrieved from the historical database by setting a time range (matching period one). The first target detection data covers abnormal behavior feature change data one (such as abnormal behavior movement amplitude, frequency changes, etc.) and abnormal behavior impact range change data one (such as the expansion of the spatial range involved in abnormal behavior). These data can reflect the dynamic changes of abnormal behavior in similar time periods in the past from different angles.

[0106] The calculation is performed using the abnormal behavior feature change data 1 and the abnormal behavior influence range change data 1 in the first target detection data in combination with time information. For example, if the diffusion rate is calculated based on the change in the abnormal behavior influence range, the difference in the abnormal behavior influence range at different time points in time period 1 can be calculated, and then divided by the corresponding time interval to obtain the diffusion rate based on the change in the influence range. For the abnormal behavior feature change data 1, the diffusion rate can be comprehensively calculated by analyzing the rate of change of features such as movement amplitude and frequency over time and combining weights (set according to the importance of the feature to the diffusion effect). The data fitting method can also be used to perform curve fitting on the feature data at different time points, and the diffusion rate can be determined based on the slope of the fitting curve.

[0107] By processing and calculating the first target detection data, a quantified first abnormal behavior diffusion rate can be obtained.

[0108] The rate impact factor of the combined location area is collected, specifically including the following steps:

[0109] Calculate the distance between the abnormal location point and the suspected abnormal location point in the combined location area and output the distance data information;

[0110] After detecting the interference degree of the abnormal behavior connection area between the abnormal location point and the suspected abnormal location point in the combined location area, interference degree value data information is obtained;

[0111] The distance data and the interference level data are combined into a rate impact factor.

[0112] This application determines that there is a certain spatial relationship between abnormal location points and suspected abnormal location points in the combined location area. By obtaining the coordinate information of two points in the monitoring coordinate system (or the building space coordinate system), the distance calculation formula (such as the Euclidean distance formula for a two-dimensional plane) is used to calculate the distance between the two points. This calculation process is based on the principles of spatial geometry and converts the coordinate difference into an actual distance value.

[0113] Distance data reflects the spatial proximity between two locations. Generally speaking, the closer the distance, the greater the likelihood that abnormal behavior will spread from the confirmed abnormal location to the suspected abnormal location. Distance is an important factor influencing the subsequent analysis of the abnormal behavior spread rate. For example, in a fire scenario, adjacent rooms are more susceptible to fire spread.

[0114] The connection area between the abnormal location and the suspected abnormal location may contain various factors that interfere with the spread of abnormal behavior. The degree of interference is detected through sensors (such as surveillance cameras analyzing obstructions in the image, and environmental monitoring equipment detecting environmental factors such as airflow and temperature) or image analysis technology (identifying whether there are obstacles or crowds in the connection area). Different interference factors are quantified, such as categorizing the degree of obstruction as no obstruction, partial obstruction, and severe obstruction and assigning corresponding numerical values, and converting airflow intensity into a numerical indicator. Ultimately, a comprehensive interference level value is obtained.

[0115] The interference level reflects the degree to which a connected area hinders or promotes the diffusion of abnormal behavior. A low interference level facilitates the diffusion of abnormal behavior, while a high interference level inhibits diffusion. For example, in a smoke diffusion scenario, a well-ventilated, unobstructed connected area facilitates smoke diffusion, while a large number of obstacles or strong headwinds hinder it.

[0116] The distance data and interference level data obtained previously are combined according to specific rules. A weighted summation approach can be used, assigning weights to the distance data and interference level values ​​(e.g., 0.6 for distance and 0.4 for interference level) based on their importance to the diffusion rate of abnormal behavior in actual security scenarios. The rate impact factor is then calculated by weighted summation. Alternatively, the two data sets can be combined by constructing a more complex functional relationship (e.g., a mapping relationship based on machine learning model training).

[0117] The rate impact factor comprehensively considers spatial distance and environmental interference factors, providing a comprehensive adjustment parameter for subsequent analysis of the diffusion rate of suspected anomaly locations. When calculating diffusion rate ratio data, the rate impact factor can make the results more realistic and accurately reflect the diffusion characteristics of anomaly behavior within the combined location area.

[0118] Processing is performed based on the correlation degree, the diffusion rate of the first abnormal behavior, and the rate influencing factor to obtain diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location point, specifically including the following steps:

[0119] The abnormal behavior diffusion rate of the suspected abnormal location point in the combined location area is predicted according to the correlation degree, the first abnormal behavior diffusion rate and the rate influencing factor to obtain a second predicted diffusion rate;

[0120] Predicting a third predicted diffusion rate of a suspected abnormal location point in the combined location area based on the first target detection data;

[0121] Processing the historical period detection data to obtain a fourth predicted diffusion rate of the suspected abnormal location point in the combined location area;

[0122] calculating a ratio between the second predicted diffusion rate and the third predicted diffusion rate to obtain diffusion rate ratio data one;

[0123] A ratio between the third predicted diffusion rate and the fourth predicted diffusion rate is calculated to obtain diffusion rate ratio data two.

[0124] The correlation degree of this application reflects the degree of correlation between the determined abnormal behavior location point and the suspected abnormal location point. The first abnormal behavior diffusion rate reflects the diffusion speed of the currently determined abnormal point, and the rate influencing factor integrates the spatial distance and interference factors. These three elements are input into a pre-established prediction model (which can be a regression model based on statistics, such as multiple linear regression, or a machine learning model, such as an artificial neural network). The model predicts the abnormal behavior diffusion rate of the suspected abnormal location point in the combined location area by learning the relationship between these elements and the abnormal behavior diffusion rate, and obtains the second predicted diffusion rate. For example, in the multiple linear regression model, let the correlation degree be x1, the first abnormal behavior diffusion rate be x2, and the rate influencing factor be x3. By training the coefficients a1, a2, and a3, the prediction formula y=a1x1+a2x2+a3x3+b is obtained, where y is the second predicted diffusion rate.

[0125] The predicted rate is based on current real-time data and related influencing factors, and can reflect the abnormal behavior diffusion trend of suspected abnormal locations under the current conditions.

[0126] The first target detection data includes data on changes in abnormal behavior characteristics and changes in the impact range of abnormal behavior within a historical period. Using time series analysis methods (such as the ARIMA model) or empirical models based on historical data, the diffusion rate of suspected abnormal locations within the combined location area is predicted based on the changing patterns of these data. For example, in the ARIMA model, the relevant indicators in the first target detection data are arranged in chronological order. The model parameters are determined by analyzing the autocorrelation and partial autocorrelation of the historical data, and the diffusion rate for the future period is then predicted, i.e., the third predicted diffusion rate.

[0127] This predicted rate provides a reference value based on the development law of past abnormal behavior from the perspective of similar historical periods, which is used to compare and analyze the rationality and changing trend of the diffusion rate of the current suspected abnormal location point.

[0128] Perform more comprehensive processing of historical detection data. More relevant feature data can be extracted from different historical periods (e.g., abnormal behavior data from different seasons and time periods). Data mining techniques (e.g., cluster analysis, decision trees) or complex time series models (e.g., LSTM neural networks) can be used to uncover potential patterns and regularities within the data, thereby predicting the fourth-order diffusion rate of suspected abnormal locations within the combined location area. For example, LSTM neural networks can predict future diffusion rates by memorizing long-term dependencies in historical data.

[0129] The fourth predicted diffusion rate integrates a wider range of historical information and can estimate the diffusion rate of suspected abnormal locations from a macro historical perspective, providing another dimension of reference data for subsequent ratio calculations, making the analysis results more reliable and comprehensive.

[0130] Diffusion rate ratio data 1 is calculated by calculating the ratio of the second predicted diffusion rate to the third predicted diffusion rate: ratio data 1 = second predicted diffusion rate / third predicted diffusion rate. It reflects the relative relationship between the diffusion rate predicted based on current real-time data and the diffusion rate predicted based on data from a similar historical period. Diffusion rate ratio data 2 is calculated by calculating the ratio of the third predicted diffusion rate to the fourth predicted diffusion rate: ratio data 2 = third predicted diffusion rate / fourth predicted diffusion rate. It reflects the comparison between the diffusion rate predicted based on data from a similar historical period and the diffusion rate predicted based on a broader historical data set.

[0131] These two ratios measure the changes and relative trends in the diffusion rate of suspected anomaly locations from different perspectives. Analyzing these two ratios allows us to determine whether the current diffusion rate of a suspected anomaly location conforms to historical patterns or exhibits unusual changes, providing key quantitative evidence for ultimately determining whether the area is anomalous.

[0132] Processing the historical period detection data to obtain a fourth predicted diffusion rate of the suspected abnormal location point in the combined location area specifically includes the following steps:

[0133] Extracting second target detection data within period 2 from the historical period detection data; wherein the second target detection data includes abnormal behavior feature change data 2 and abnormal behavior impact range change data 2;

[0134] A fourth predicted diffusion rate of the suspected abnormal position point in the combined position area is predicted based on the second target detection data.

[0135] Historical period detection data stores information related to abnormal behavior at different times in the past. To obtain historical data relevant to the current analysis, it is necessary to filter out the second target detection data within period two from this data. By setting a specific time range (period two), a search is performed in the historical database. The second target detection data includes abnormal behavior feature change data 2 (for example, changes in the abnormal behavior's movement pattern, frequency, etc. during period two) and abnormal behavior impact range change data 2 (for example, changes in the spatial range involved in the abnormal behavior during period two).

[0136] The extracted secondary target detection data provides a concrete historical basis for subsequent predictions of the diffusion rate of suspected anomaly locations. Focusing on the details of abnormal behavior during a specific historical period helps analysts understand the development characteristics of abnormal behavior under similar circumstances in the past.

[0137] The second target detection data is processed using methods such as time series analysis, machine learning, or statistical modeling. For example, exponential smoothing in time series analysis can be used to predict the future diffusion rate (i.e., the suspected abnormal location point in the current combined location area) based on the temporal trends of the abnormal behavior feature change data 2 and the abnormal behavior impact range change data 2. Alternatively, a regression algorithm in machine learning can be used to train a model using the various features in the second target detection data as independent variables and the past diffusion rate as the dependent variable. The trained model can then be used to predict the fourth predicted diffusion rate of the current suspected abnormal location point.

[0138] The fourth predicted diffusion rate, derived from analyzing the second target detection data, provides an estimate of the diffusion of the suspected anomaly point from another historical perspective. Comparing this value with other predicted diffusion rates (such as the third predicted diffusion rate) helps determine whether the current diffusion trend of the suspected anomaly point conforms to historical patterns, and furthermore, whether the area within the combined location area containing the suspected anomaly point is an area of ​​abnormal behavior.

[0139] The diffusion rate ratio data 1 and the diffusion rate ratio data 2 are processed and analyzed to determine whether the area where the suspected abnormal location point in the combined location area is located is an abnormal behavior location area, specifically including the following steps:

[0140] The change ratio between the diffusion rate ratio data 1 and the diffusion rate ratio data 2 is calculated to obtain a change ratio value;

[0141] Preset a diffusion rate change ratio threshold, and compare the change ratio value with the preset diffusion rate change ratio threshold;

[0142] If the change ratio value is greater than the preset diffusion rate change ratio threshold, the suspected abnormal location point in the combined location area is determined to be an abnormal behavior location area;

[0143] If the change ratio value is less than or equal to the preset diffusion rate change ratio threshold, the suspected abnormal location point in the combined location area is determined to be a non-abnormal behavior location area.

[0144] Diffusion rate ratio data 1 and 2 in this application reflect the relative relationship between the diffusion rates of suspected anomaly locations from different perspectives. The change ratio between the two is calculated. This calculation is based on the principle of proportionality, and the magnitude of the change between the two ratio data is quantified by comparing the degree of difference between them.

[0145] The change ratio value intuitively reflects the relative change between Diffusion Rate Ratio Data 1 and Diffusion Rate Ratio Data 2, and is a key quantitative indicator for subsequent judgment. It reflects the degree of fluctuation in the diffusion rate of the suspected anomaly location under different reference dimensions.

[0146] The preset diffusion rate change ratio threshold is a standard value set based on extensive historical data and actual security experience. Comparing the calculated change ratio value with this threshold is based on statistical and empirical logic. If the change ratio value is greater than the preset threshold, it indicates that the diffusion rate of the suspected abnormal location point has changed beyond the normal fluctuation range in different reference dimensions, and its diffusion trend has undergone abnormal changes, indicating that it is an abnormal behavior location area. If the change ratio value is less than or equal to the preset threshold, the diffusion rate change is within the normal range, and the area where the suspected abnormal location point is located is not an abnormal behavior location area.

[0147] Through this comparison and judgment mechanism, it is possible to make a clear judgment on whether the area where the suspected abnormal location point in the combined location area is located is an abnormal behavior location area.

[0148] The following steps are also included:

[0149] Detect and count the environmental change values ​​of the building security monitoring area during time periods 1, 2, and 3 for suspected abnormal location points in the combined location area based on the non-abnormal behavior location area;

[0150] Preset an environmental change degree warning value, and compare the environmental change degree value with the preset environmental change degree warning value;

[0151] If the environmental change degree value is greater than or equal to the preset environmental change degree warning value, the abnormal behavior risk warning result is output;

[0152] If the environmental change degree value is less than the preset environmental change degree warning value, the abnormal behavior risk warning result is output.

[0153] This application continues to monitor the environmental changes in the building security monitoring area within time periods one, two, and three at suspected abnormal location points in the combined location area after determining that the location area is not abnormal. Environmental data is collected through various sensors (such as temperature and humidity sensors, light sensors, air quality sensors, etc.) and image analysis technology (such as analyzing the flow of people and objects in the monitoring screen). The collected data is processed and analyzed, for example, the change range of environmental parameters (such as temperature, humidity, light intensity, etc.) in different time periods is calculated, and information such as the movement frequency and trajectory changes of people or objects are counted, and then these data are integrated to obtain a quantitative value of the degree of environmental change.

[0154] The environmental change degree value comprehensively reflects the dynamic changes in the environment of the suspected abnormal location within a specific period of time. It can capture changes in environmental factors that may affect the occurrence or development of abnormal behavior, providing new dimensions of information for further assessing abnormal behavior risks.

[0155] The preset environmental change degree warning value is a standard value pre-set based on the normal environmental fluctuation range of the building security monitoring area and the correlation between historical abnormal events and environmental changes. The statistically obtained environmental change degree value is compared with the preset warning value, and the risk assessment logic is used to determine whether there is a potential risk of abnormal behavior. If the environmental change degree value is greater than or equal to the preset warning value, it means that the current environmental change exceeds the normal range and may cause abnormal behavior. Therefore, an abnormal behavior risk warning result is output to prompt relevant personnel to pay attention and take preventive measures. If the environmental change degree value is less than the preset warning value, it indicates that the environmental change is within the normal range and there is no obvious risk of abnormal behavior due to environmental changes. The corresponding warning result is also output (i.e., it indicates that there is no obvious risk).

[0156] The device embodiments described above are merely illustrative. 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 units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0157] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0158] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. An intelligent building security monitoring system, characterized in that: include: Acquisition and extraction module: collects video image information of the building security monitoring area, extracts feature data of different degrees from video image information with abnormal behavior to obtain a first feature video image and a second feature video image; A division module: divides the second feature video image into a target video image and a locked video image according to the abnormal behavior in the second feature video image; wherein the target video image is a video image with abnormal behavior, and the locked video image is a video image suspected of having abnormal behavior; Image extraction module: extracts the attention video image corresponding to the target video image from the first feature video image, and extracts the positioning video image corresponding to the locked video image from the first feature video image; Combination module: combines the abnormal behavior location points determined in the video image of interest with the suspected abnormal location points in the positioning video image to obtain a combined location area; Processing and analysis module: processes and analyzes the abnormal behavior location points and suspected abnormal location points in the combined location area to obtain diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location points; Judgment module: Process and analyze the diffusion rate ratio data 1 and the diffusion rate ratio data 2 to determine whether the area where the suspected abnormal location point in the combined location area is located is an abnormal behavior location area.

2. The intelligent building security monitoring system according to claim 1, characterized in that: Processing and analyzing the abnormal behavior location points and the suspected abnormal location points in the combined location area to obtain diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location points specifically includes the following steps: The correlation degree is obtained by extracting the correlation degree between the abnormal behavior location point and the suspected abnormal location point in the combined location area; The abnormal behavior diffusion rate within the time period 1 of the abnormal behavior location point determined in the combined location area is statistically analyzed to obtain a first abnormal behavior diffusion rate; The rate impact factor of the combined location area is collected, and the diffusion rate ratio data 1 and the diffusion rate ratio data 2 of the suspected abnormal location point are obtained by processing according to the correlation degree, the diffusion rate of the first abnormal behavior and the rate impact factor.

3. The intelligent building security monitoring system according to claim 2, characterized in that: The method of obtaining a first abnormal behavior diffusion rate by statistically analyzing the abnormal behavior diffusion rate within a time period 1 of the abnormal behavior location point determined in the combined location area specifically includes the following steps: Compare the correlation degree with a preset correlation threshold, and obtain historical period detection data when the correlation degree is greater than or equal to the preset correlation threshold; The historical period detection data and the abnormal behavior diffusion rate of the abnormal location point determined in the combined location area are processed to obtain a first abnormal behavior diffusion rate.

4. The intelligent building security monitoring system according to claim 3, characterized in that: The abnormal behavior diffusion rate of the abnormal location point determined in the historical period detection data and the combined location area is processed to obtain a first abnormal behavior diffusion rate, specifically comprising the following steps: Extracting first target detection data within a time period 1 from the historical time period detection data; wherein the first target detection data includes abnormal behavior feature change data 1 and abnormal behavior impact range change data 1; The abnormal behavior diffusion rate of the abnormal position point determined in the combined position area is statistically analyzed according to the first target detection data to obtain a first abnormal behavior diffusion rate.

5. The intelligent building security monitoring system according to claim 4, characterized in that: The rate impact factor of the combined location area is collected, specifically including the following steps: Calculate the distance between the abnormal location point and the suspected abnormal location point in the combined location area and output the distance data information; Detecting interference degree value data information of the abnormal behavior connection area between the determined abnormal location point and the suspected abnormal location point in the combined location area; The distance data information and the interference level value data information are combined into a rate impact factor.

6. The intelligent building security monitoring system according to claim 5, characterized in that: Processing is performed based on the correlation degree, the diffusion rate of the first abnormal behavior, and the rate influencing factor to obtain diffusion rate ratio data 1 and diffusion rate ratio data 2 of the suspected abnormal location point, specifically including the following steps: The abnormal behavior diffusion rate of the suspected abnormal location point in the combined location area is predicted according to the correlation degree, the first abnormal behavior diffusion rate and the rate influencing factor to obtain a second predicted diffusion rate; Predicting a third predicted diffusion rate of a suspected abnormal location point in the combined location area based on the first target detection data; Processing the historical period detection data to obtain a fourth predicted diffusion rate of the suspected abnormal location point in the combined location area; calculating a ratio between the second predicted diffusion rate and the third predicted diffusion rate to obtain diffusion rate ratio data one; The ratio between the third predicted diffusion rate and the fourth predicted diffusion rate is calculated to obtain diffusion rate ratio data 2.

7. The intelligent building security monitoring system according to claim 6, characterized in that: Processing the historical period detection data to obtain a fourth predicted diffusion rate of the suspected abnormal location point in the combined location area specifically includes the following steps: Extracting second target detection data within period 2 from the historical period detection data; wherein the second target detection data includes abnormal behavior feature change data 2 and abnormal behavior impact range change data 2; A fourth predicted diffusion rate of the suspected abnormal position point in the combined position area is predicted based on the second target detection data.

8. The intelligent building security monitoring system according to claim 7, characterized in that: The diffusion rate ratio data 1 and the diffusion rate ratio data 2 are processed and analyzed to determine whether the area where the suspected abnormal location point in the combined location area is located is an abnormal behavior location area, specifically including the following steps: The change ratio between the diffusion rate ratio data 1 and the diffusion rate ratio data 2 is calculated to obtain a change ratio value; Preset a diffusion rate change ratio threshold, and compare the change ratio value with the preset diffusion rate change ratio threshold; If the change ratio value is greater than the preset diffusion rate change ratio threshold, the suspected abnormal location point in the combined location area is determined to be an abnormal behavior location area; If the change ratio value is less than or equal to the preset diffusion rate change ratio threshold, the suspected abnormal position point in the combined position area is determined to be a non-abnormal behavior position area.

9. The intelligent building security monitoring system according to claim 8, characterized in that: The following steps are also included: Detect and count the environmental change values ​​of the building security monitoring area during time periods 1, 2, and 3 for suspected abnormal location points in the combined location area based on the non-abnormal behavior location area; Preset an environmental change degree warning value, and compare the environmental change degree value with the preset environmental change degree warning value; If the environmental change degree value is greater than or equal to the preset environmental change degree warning value, the abnormal behavior risk warning result is output; If the environmental change degree value is less than the preset environmental change degree warning value, the abnormal behavior risk warning result is output.

Images