A smart community video analysis system based on big data

Abstract

The present invention discloses a smart community video analysis system based on big data, comprising a data acquisition module, a data processing module, and a data output module; the data acquisition module is used for collecting real-time data information of target objects in the community; the data processing module is used for receiving real-time data information and standard data information, processing and analyzing the real-time data information in combination with the standard data information, obtaining the potential risks under the behavior of the target object, and inferring the level of the potential risk of the target object, and generating an early warning report; the present invention obtains information such as the size, path, and speed of the target object, analyzes the probability of a collision of the target object in a preset area, and classifies the risk level of objects that may collide according to the collision probability, and issues an early warning to remind the target object to respond in time, which can effectively reduce the possibility of a collision of the target object, thereby improving the safety of the living environment in the community.

Description

Technical Field

[0001] The present invention relates to the field of video data analysis, and in particular to a smart community video analysis system based on big data. Background Art

[0002] With the development and progress of smart communities, video surveillance systems play a vital role in maintaining community security and management. The security of smart communities has always been a top priority in management work, especially road driving safety, which has always been an urgent problem to be solved. With the development of electronic products and changes in lifestyles, a large number of people who are addicted to their phones have emerged, paying attention to their electronic products all the time. At the same time, a large number of people wear headphones when traveling. At the same time, the safety awareness of the elderly and children in the community is not high; and now the noise of cars is getting smaller and smaller, accompanied by blind spots in vision, a large number of safety accidents occur in the community every year, causing losses of resources and property.

[0003] The current management method mainly relies on the video surveillance system, but video surveillance can only be used as a basis for determining responsibility, and cannot provide early warning to prevent safety accidents. In order to further improve the safety issues in the community, a smart community video analysis system based on big data is proposed. Summary of the invention

[0004] The technical problem to be solved by the present invention is: how to solve the problem that the existing security protection system has a single protection type, resulting in poor protection effect and bringing certain impacts on the use of the security protection system, and provides a smart community video analysis system based on big data.

[0005] The present invention solves the above technical problems through the following technical solutions. The present invention includes a data acquisition module, a preprocessing module, a big data module, a data processing module, and a data output module;

[0006] The data collection module is used to collect real-time data information of target objects in the community;

[0007] The preprocessing module is used to detect real-time data information, and to denoise and enhance the real-time data information;

[0008] The big data module is used to receive real-time data information, and process, manage and analyze the real-time data information to obtain standard data information of the target object;

[0009] The data processing module is used to receive real-time data information and standard data information, process and analyze the real-time data information in combination with the standard data information, obtain the potential risks under the behavior of the target object, infer the level of the potential risk of the target object, and generate an early warning report;

[0010] The data output module is used to receive and output early warning reports.

[0011] Preferably, the real-time data information includes real-time size data and real-time speed data, the standard data information includes standard path information of the target object's moving trajectory, the potential risk includes abnormal collision events, and the specific analysis process of the data processing module is as follows:

[0012] Obtain the real-time size data of the target object and extract the height size H of the target object;

[0013] When the height dimension H of the target object is less than or equal to the preset threshold Z, the data is discarded;

[0014] When the height dimension H of the target object is greater than the preset threshold Z, a circular area Y1 is established with the geometric center of the target object as the center and the preset dimension R as the radius;

[0015] Get the real-time velocity data of two target objects in the circular area Y1, which are V1 and V2 respectively;

[0016] Import standard path information, obtain the moving paths between two target objects in the standard path information, and obtain the path length dimension Dx between the two target objects;

[0017] Use the first rule to calculate the relative velocity V' between the two target objects;

[0018] The first rule is:

[0019] Select any point between the two target objects as the reference point A, and obtain the real-time size data of the two target objects from the reference point A, which are d1 and d2 respectively, and d1>0, d2>0;

[0020] At the next time point, the real-time size data of the two target objects from the reference point A are obtained again, which are d1` and d2` respectively, and d1`>0, d2`>0;

[0021] Calculate the change in size Δdx=dx`-dx;

[0022] If Δd1≤0, Δd2≤0, then V`=V1+V2; if Δd1≥0, Δd2≥0, then the data is eliminated;

[0023] If Δd1≥0, Δd2≤0, then V`=V2-V1, when V`<0, the data is eliminated; if Δd1≤0, Δd2≥0, then V`=V1-V2, when V`<0, the data is eliminated;

[0024] Use the formula T=Dx / V` to calculate the encounter time T of the target object;

[0025] When the encounter time T> the preset threshold Q1, the data is eliminated;

[0026] When the encounter time T is less than the preset threshold Q2, it indicates that the potential risk level is high and a first-level warning report is generated;

[0027] When the preset threshold Q2 ≤ encounter time T ≤ preset threshold Q1, it indicates that the potential risk level is low, and a level 2 warning report is generated.

[0028] Preferably, the calculation process of the relative speed V' also includes the following process:

[0029] Continuously obtain the average speeds Vr1, Vr2, Vr3, ..., Vrx of multiple unit time periods within a period of time when the target object enters the circular area Y;

[0030] Using the formula Calculate the average speed of multiple unit time periods Vr1, Vr2, Vr3, ..., Vrx ;

[0031] Randomly select the average speed of n time periods in multiple unit time periods, n>0, and calculate the deviation degree P of the speed in each time period. The calculation process is as follows:

[0032] ;

[0033] When P ≤ the preset threshold W, it means that the moving speed of the target object is balanced, so let ;

[0034] When P>preset threshold W, it means that the moving speed of the target object fluctuates greatly, and the following processing is performed:

[0035] Calculate the aggregation state β of the average speeds Vr1, Vr2, Vr3, ..., Vrx of multiple unit time periods. The specific calculation process is:

[0036] ;

[0037] make , obtain the average speed Vrx of β in this interval, and calculate the mean speed of the average speed Vrx of β in this interval , then let .

[0038] Preferably, the system further includes a database module, the database module is used to store historical data information of the movement of the target object, the historical data information includes the travel data information of the target object of the same type, and the establishment process of the circular area Y1 also includes:

[0039] Obtain the driving data information of the target object, obtain the time t1 required for the target object to decelerate from the speed Vx to 0, and let the human body's reaction time be t2;

[0040] Calculate the radius size R` of the circular area Y;

[0041] When the target object is a single independent individual, the specific calculation process is:

[0042] ;

[0043] Let the preset size R = R`;

[0044] When the target object is a plurality of independent entities, the real-time size data of the target object within a period of time is continuously obtained, which are L1, L2, L3, ..., Lx;

[0045] Calculate the dimensional stability coefficient K of the target object. The specific calculation process is as follows:

[0046] Using the formula Calculate the average size of the target object ;

[0047] Calculate the stability coefficient K. The specific calculation process is:

[0048] ;

[0049] When the stability coefficient K≤the preset threshold U, the preset size R=R`;

[0050] When the stability coefficient K> the preset threshold U, the preset size .

[0051] Preferably, when the encounter time T is less than the preset threshold value Q2, the data processing module further performs the following processing:

[0052] Establish a virtual path, which is a straight line, and the two target objects are both within the virtual path, so that the two target objects move in the same or opposite directions, so that the positions of the two target objects in the width direction of the virtual path are consistent with the positions of the target objects on the actual path, and the path length dimension between the two target objects is Dx;

[0053] Establish two circumscribed circles of the target object, namely circumscribed circle Y2 and circumscribed circle Y3, and obtain the radius size of circumscribed circle Y2 as Dr1 and the radius size of circumscribed circle Y3 as Dr2;

[0054] Establish straight lines Ln1 and Ln2, both of which are tangent to circumscribed circles Y2 and Y3, and intersect at point B;

[0055] Calculate the maximum distance between two target objects when they collide , the specific calculation process is:

[0056] ;

[0057] Wherein, α is the angle between any straight line Ln1 and straight line Ln2 and the geometric center line of the two target objects when the straight line is parallel to the straight line where the target object moves;

[0058] Obtain the positional relationship between the two target objects on the actual path, and use the above method to transform the two target objects into the virtual path, and obtain the center distance L' of the two target objects;

[0059] when When the potential risk level is changed, the first-level warning report is changed to a second-level warning report.

[0060] Preferably, the process of establishing the circumscribed circle Y2 and the circumscribed circle Y3 further includes the following processing:

[0061] Import standard path information of target object's moving trajectory;

[0062] In the process of establishing the circumscribed circle Y2, the increment J1 of the moving path of the target object in the direction perpendicular to its moving direction is obtained, and the radius size of the circumscribed circle Y2 is changed to Dr3, and Dr3=Dr1+J1;

[0063] During the process of establishing the circumscribed circle Y3, the increment J2 of the target object's moving path perpendicular to its moving direction is obtained, and the radius of the circumscribed circle Y3 is changed to Dr4, and Dr4=Dr1+J2.

[0064] Preferably, when the preset threshold Q2≤the encounter time T≤the preset threshold Q1, the data processing module further performs the following processing:

[0065] Importing real-time size data of two target objects respectively, wherein the real-time size data includes the length Li, width Di, and height Hi of the target objects;

[0066] Use the formula Vix=Li·Di·Hi to calculate the volumes of the two target objects, which are Vi1 and Vi2 respectively;

[0067] Calculate the volume difference S of the two target objects. The specific calculation process is:

[0068] ;

[0069] Import the real-time speeds of two target objects, which are V1 and V2 respectively;

[0070] Calculate the speed difference G of the two target objects. The specific calculation process is:

[0071] ;

[0072] Assign a correction value e1 to the volume difference S, and a correction value e2 to the speed difference G, and e1+e2=1, e2>e1, and estimate the risk factor f when the target object collides. The specific estimation process is: ;

[0073] When the risk factor f ≥ the preset threshold C, the potential risk level is changed and the second-level warning report is changed to a first-level warning report.

[0074] Preferably, the process of calculating the volume of the target object further includes the following processing:

[0075] For the same target object, its length Li, width Di, and height Hi are obtained at multiple points respectively, and the multiple points are evenly arranged, and the average value of the length Li, the average value of the width Di, and the average value of the height Hi are calculated respectively. When calculating the volume Vi, the average value of the length Li, the average value of the width Di, and the average value of the height Hi are taken respectively.

[0076] Compared with the prior art, the present invention has the following advantages: when the system is in use, it selects objects that may collide, and by obtaining information such as the size, path and speed of the target objects, it analyzes the probability of the target objects colliding in a preset area, and classifies the risk levels of the objects that may collide according to the collision probability, and issues early warnings to remind the target objects to respond in time, which can effectively reduce the possibility of collision of the target objects, thereby improving the safety of the living environment in the community, while also improving the quality of life of residents in the community, making the system more worthy of promotion and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is a schematic diagram of the overall process of the present invention. DETAILED DESCRIPTION

[0078] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation method and a specific operation process are given, but the protection scope of the present invention is not limited to the following embodiment.

[0079] like Figure 1 As shown, this embodiment provides a technical solution: a smart community video analysis system based on big data, including a data acquisition module, a preprocessing module, a big data module, a data processing module, and a data output module;

[0080] The data collection module is used to collect real-time data information of target objects in the community;

[0081] The preprocessing module is used to detect real-time data information and to denoise and enhance the real-time data information;

[0082] It should be noted that the preprocessing module processes the real-time data information in the following steps:

[0083] S1. Noise Removal: Use noise reduction algorithms, such as filter-based methods such as mean filtering, median filtering, or more advanced algorithms such as bilateral filtering or deep learning noise reduction models, to remove noise from the picture and improve detail visibility;

[0084] S2. Sharpening: By applying sharpening filters, such as Unsharp Mask or edge detection-based sharpening technology, the edges of the picture are made more obvious and look clearer;

[0085] S3, super-resolution: using interpolation or machine learning methods, such as bilinear interpolation, bicubic or deep learning SR algorithms, to enhance low-resolution videos to high resolution;

[0086] S4. Deblocking: For compressed videos, there may be obvious blocky textures. Deblocking technology can help reduce these jagged effects and make the picture look more natural.

[0087] S5. Color correction: adjust the color level, contrast and saturation to ensure the color is accurate;

[0088] S6. Compression damage repair: For quality loss caused by compression, a certain degree of recovery can be achieved through specific algorithms;

[0089] S7. Video encoding optimization: Select appropriate encoder settings, appropriately reduce the bit rate but maintain high-quality encoding methods can also improve video clarity;

[0090] Noise reduction and enhancement of video data are existing technologies, and the present invention is not limited thereto. Any technology that can reduce noise and enhance video data can be used; and video noise reduction and enhancement are not the core content of the present invention and will not be described in detail here.

[0091] The big data module is used to receive real-time data information, and process, manage and analyze the real-time data information to obtain standard data information of the target object;

[0092] The data processing module is used to receive real-time data information and standard data information, process and analyze the real-time data information in combination with the standard data information, obtain the potential risks under the behavior of the target object, infer the level of the potential risk of the target object, and generate an early warning report;

[0093] The data output module is used to receive and output early warning reports.

[0094] It should be noted that the big data module refers to the technology used in the system to process, manage and analyze large-scale data sets. It mainly includes the following key parts:

[0095] Data collection: obtaining real-time or historical data from various data sources;

[0096] Data storage: Using a distributed storage system, PB-level data can be efficiently stored and managed;

[0097] Data processing: Execute complex computing tasks through batch or stream processing engines;

[0098] Data analysis: Use statistical analysis and machine learning algorithms, such as Hadoop Mahout or Spark MLlib, to discover patterns and insights in data;

[0099] Data visualization: presenting analysis results in the form of graphs or charts to facilitate understanding and decision-making;

[0100] Data governance: Ensure the quality, consistency and security of data, including data cleaning, metadata management, etc.

[0101] The big data module can learn the living habits and movement patterns of similar individuals in the community. This part is the existing technology, and the big data module in the present invention is not limited to the data processing process of the above-mentioned parts. As long as the data processing and learning process can be completed and the analysis results can be presented, it is not the core technical content of the present invention and will not be repeated here.

[0102] In a further embodiment, the real-time data information includes real-time size data and real-time speed data, the standard data information includes standard path information of the target object's moving trajectory, and the potential risk includes abnormal collision events. The specific analysis process of the data processing module is as follows:

[0103] Obtain the real-time size data of the target object and extract the height size H of the target object;

[0104] When the height dimension H of the target object is less than or equal to the preset threshold Z, the data is discarded;

[0105] When the height dimension H of the target object is greater than the preset threshold value Z, a circular area Y1 is established with the geometric center of the target object as the center and the preset dimension R as the radius;

[0106] Get the real-time velocity data of two target objects in the circular area Y1, which are V1 and V2 respectively;

[0107] Import standard path information, obtain the moving paths between two target objects in the standard path information, and obtain the path length dimension Dx between the two target objects;

[0108] The standard path information is the predicted trajectory made by the big data module based on the movement trajectories of a large number of objects of the same type;

[0109] Use the first rule to calculate the relative velocity V' between the two target objects;

[0110] The first rule is:

[0111] Select any point between the two target objects as the reference point A, and obtain the real-time size data of the two target objects from the reference point A, which are d1 and d2 respectively, and d1>0, d2>0;

[0112] At the next time point, the real-time size data of the two target objects from the reference point A are obtained again, which are d1` and d2` respectively, and d1`>0, d2`>0;

[0113] Calculate the change in size Δdx=dx`-dx;

[0114] If Δd1≤0, Δd2≤0, then V`=V1+V2; if Δd1≥0, Δd2≥0, then the data is eliminated;

[0115] If Δd1≥0, Δd2≤0, then V`=V2-V1, when V`<0, the data is eliminated; if Δd1≤0, Δd2≥0, then V`=V1-V2, when V`<0, the data is eliminated;

[0116] Use the formula T=Dx / V` to calculate the encounter time T of the target object;

[0117] When the encounter time T> the preset threshold Q1, the data is eliminated;

[0118] When the encounter time T is less than the preset threshold Q2, it indicates that the potential risk level is high, and a first-level warning report is generated. A relatively strong external warning method can be applied to the first-level warning report, such as a sound alarm plus a red flashing light;

[0119] When the preset threshold Q2≤encounter time T≤preset threshold Q1, it indicates that the potential risk level is low, and a level 2 warning report is generated. The level 2 warning report can impose a relatively mild external warning method, such as a yellow flashing light.

[0120] In order to prevent the early warning from disturbing the public, different circular areas Y1 are set for different positions and speeds, and effective reaction time is reserved for target objects that may collide, so that they have sufficient time to react in the event of a possible collision and issue an early warning to the target objects that may collide; at the same time, it prevents the early warning from being triggered frequently and interfering with the normal lives of community residents, thereby improving the safety of the community while ensuring the comfort of the community environment.

[0121] Furthermore, the calculation process of the relative speed V' also includes the following process:

[0122] Continuously obtain the average speeds Vr1, Vr2, Vr3, ..., Vrx of multiple unit time periods within a period of time when the target object enters the circular area Y;

[0123] Using the formula Calculate the average speed of multiple unit time periods Vr1, Vr2, Vr3, ..., Vrx ;

[0124] It should be noted that the average speed in multiple unit time periods is the average of the real-time speed at the start point and the real-time speed at the end point of the time period;

[0125] Randomly select the average speed of n time periods in multiple unit time periods, n>0, and calculate the deviation degree P of the speed in each time period. The calculation process is as follows:

[0126] ;

[0127] When P ≤ the preset threshold W, it means that the moving speed of the target object is balanced, so let ;

[0128] When P>preset threshold W, it means that the moving speed of the target object fluctuates greatly, and the following processing is performed:

[0129] Calculate the aggregation state β of the average speeds Vr1, Vr2, Vr3, ..., Vrx of multiple unit time periods. The specific calculation process is:

[0130] ;

[0131] make , obtain the average speed Vrx of β in this interval, and calculate the mean speed of the average speed Vrx of β in this interval , then let .

[0132] There are a certain number of animals and children in the community. Their safety awareness is weak and they cannot judge the potential risks of specific behaviors. Therefore, their behavior patterns fluctuate greatly and may change randomly. In order to further prevent collisions, the special conditions of independent individuals are analyzed to reserve longer response time and further improve the safety quality of the community.

[0133] Furthermore, the system also includes a database module, which is used to store historical data information of the movement of the target object, and the historical data information includes the travel data information of the target object of the same type. The process of establishing the circular area Y1 also includes:

[0134] Obtain the driving data information of the target object, obtain the time t1 required for the target object to decelerate from the speed Vx to 0, and let the human body's reaction time be t2;

[0135] Calculate the radius size R` of the circular area Y;

[0136] When the target object is a single independent individual, the specific calculation process is:

[0137] ;

[0138] Let the preset size R = R`;

[0139] When the target object is a plurality of independent entities, the real-time size data of the target object within a period of time is continuously obtained, which are L1, L2, L3, ..., Lx;

[0140] Calculate the dimensional stability coefficient K of the target object. The specific calculation process is as follows:

[0141] Using the formula Calculate the average size of the target object ;

[0142] Calculate the stability coefficient K. The specific calculation process is:

[0143] ;

[0144] When the stability coefficient K≤the preset threshold U, the preset size R=R`;

[0145] When the stability coefficient K> the preset threshold U, the preset size .

[0146] Here, the type of target object is analyzed to determine whether the multiple individuals move in unison when the target object is composed of multiple independent individuals. When there is a large deviation in the movement of multiple individuals, it indicates that there are individuals with irregular movement in the target object. Adaptive adjustments are made to the circular area Y1 to improve the reliability of the system operation.

[0147] When judging the possibility of collision between two target objects, the usual method is to predict the moving trajectory of the target object, then select a reference object, and judge whether the moving trajectories of the two target objects overlap. When the moving trajectories of the two target objects overlap, the two target objects may collide. This method usually selects multiple reference points on the reference object. When collecting data through video, since there may be occluders at different reference points, the occluders may be misjudged as reference objects and reference points may be selected, which will cause a large deviation in the data and affect the analysis results. At the same time, the change in the shape of the reference object itself will also cause errors in the selection of reference points and affect the analysis results. To solve this problem, the following technical solutions are proposed:

[0148] Preferably, when the encounter time T is less than the preset threshold value Q2, the data processing module further performs the following processing:

[0149] Establish a virtual path, which is a straight line, and both target objects are within the virtual path. The moving directions of the two target objects are the same or opposite, and the positions of the two target objects in the width direction of the virtual path are consistent with the positions of the target objects on the actual path, and the path length dimension between the two target objects is Dx;

[0150] It should be noted that when an object moves normally, it will choose the roadside line as a reference, and then choose an appropriate distance from the roadside line to move, and it will be relatively stable in a very short period of time; when the moving directions of two target objects are different, the two target objects can be placed in the same path for comparison. As long as the distance from the roadside line is determined, effective comparison data can be made.

[0151] Establish two circumscribed circles of the target object, namely circumscribed circle Y2 and circumscribed circle Y3, and obtain the radius size of circumscribed circle Y2 as Dr1 and the radius size of circumscribed circle Y3 as Dr2;

[0152] Establish straight lines Ln1 and Ln2, both of which are tangent to circumscribed circles Y2 and Y3, and intersect at point B;

[0153] Calculate the maximum distance between two target objects when they collide , the specific calculation process is:

[0154] ;

[0155] Wherein, α is the angle between any straight line Ln1 and straight line Ln2 and the geometric center line of the two target objects when the straight line is parallel to the straight line where the target object moves;

[0156] Obtain the positional relationship between the two target objects on the actual path, and use the above method to transform the two target objects into the virtual path, and obtain the center distance L' of the two target objects;

[0157] when When the potential risk level is changed, the first-level warning report is changed to a second-level warning report.

[0158] This preferred solution estimates the collision probability based on the target object's own data parameters, avoiding misjudgment caused by selecting reference objects, and breaking away from the limitations of roadside lines, effectively improving the accuracy of collision prediction and the effectiveness of the system's use.

[0159] The process of establishing the circumscribed circle Y2 and the circumscribed circle Y3 also includes the following processing:

[0160] Import standard path information of target object's moving trajectory;

[0161] In the process of establishing the circumscribed circle Y2, the increment J1 of the moving path of the target object in the direction perpendicular to its moving direction is obtained, and the radius size of the circumscribed circle Y2 is changed to Dr3, and Dr3=Dr1+J1;

[0162] During the process of establishing the circumscribed circle Y3, the increment J2 of the target object's moving path perpendicular to its moving direction is obtained, and the radius of the circumscribed circle Y3 is changed to Dr4, and Dr4=Dr1+J2.

[0163] In special locations, such as road bends, there will be a certain turning radius that will change the moving trajectory of the target object, and this situation is inevitable. This solution analyzes the changes in the path caused by special locations to improve the practicality of the system.

[0164] When the preset threshold Q2 ≤ the encounter time T ≤ the preset threshold Q1, the data processing module further performs the following processing:

[0165] Import the real-time size data of two target objects respectively, the real-time size data includes the length Li, width Di, and height Hi of the target objects;

[0166] Use the formula Vix=Li·Di·Hi to calculate the volumes of the two target objects, which are Vi1 and Vi2 respectively;

[0167] Calculate the volume difference S of the two target objects. The specific calculation process is:

[0168] ;

[0169] Import the real-time speeds of two target objects, which are V1 and V2 respectively;

[0170] Calculate the speed difference G of the two target objects. The specific calculation process is:

[0171] ;

[0172] Assign a correction value e1 to the volume difference S, and a correction value e2 to the speed difference G, and e1+e2=1, e2>e1, and estimate the risk factor f when the target object collides. The specific estimation process is: ;

[0173] When the risk factor f ≥ the preset threshold C, the potential risk level is changed and the second-level warning report is changed to a first-level warning report.

[0174] Obtain the volume and speed of the target object, analyze the target object when it collides to prevent possible serious damage and impact, and focus on early warning and reminders to avoid major safety accidents in the community.

[0175] The calculation process of the volume of the target object also includes the following processing:

[0176] For the same target object, its length Li, width Di, and height Hi are obtained at multiple points respectively, and the multiple points are evenly arranged, and the average value of the length Li, the average value of the width Di, and the average value of the height Hi are calculated respectively. When calculating the volume Vi, the average value of the length Li, the average value of the width Di, and the average value of the height Hi are taken respectively.

[0177] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0178] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction 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 terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0179] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A smart community video analysis system based on big data, characterized in that: It includes data acquisition module, preprocessing module, big data module, data processing module and data output module; The data collection module is used to collect real-time data information of target objects in the community; The preprocessing module is used to detect real-time data information, and to denoise and enhance the real-time data information; The big data module is used to receive real-time data information, and process, manage and analyze the real-time data information to obtain standard data information of the target object; The data processing module is used to receive real-time data information and standard data information, process and analyze the real-time data information in combination with the standard data information, obtain the potential risks under the behavior of the target object, infer the level of the potential risk of the target object, and generate an early warning report; The data output module is used to receive and output early warning reports; The real-time data information includes real-time size data and real-time speed data, the standard data information includes standard path information of the target object's moving trajectory, the potential risk includes abnormal collision events, and the specific analysis process of the data processing module is as follows: Obtain the real-time size data of the target object and extract the height size H of the target object; When the height dimension H of the target object is less than or equal to the preset threshold Z, the data is discarded; When the height dimension H of the target object is greater than the preset threshold Z, a circular area Y1 is established with the geometric center of the target object as the center and the preset dimension R as the radius; Get the real-time velocity data of two target objects in the circular area Y1, which are V1 and V2 respectively; Import standard path information, obtain the moving paths between two target objects in the standard path information, and obtain the path length dimension Dx between the two target objects; Use the first rule to calculate the relative velocity V' between the two target objects; The first rule is: Select any point between the two target objects as the reference point A, and obtain the real-time size data of the two target objects from the reference point A, which are d1 and d2 respectively, and d1>0, d2>0; At the next time point, the real-time size data of the two target objects from the reference point A are obtained again, which are d1` and d2` respectively, and d1`>0, d2`>0; Calculate the change in size Δdx=dx`-dx, x=1 or 2; If Δd1≤0, Δd2≤0, then V`=V1+V2; if Δd1≥0, Δd2≥0, then the data is eliminated; If Δd1≥0, Δd2≤0, then V`=V2-V1, when V`<0, the data is eliminated; if Δd1≤0, Δd2≥0, then V`=V1-V2, when V`<0, the data is eliminated; Use the formula T=Dx / V` to calculate the encounter time T of the target object; When the encounter time T> the preset threshold Q1, the data is eliminated; When the encounter time T is less than the preset threshold Q2, it indicates that the potential risk level is high and a first-level warning report is generated; When the preset threshold Q2 ≤ encounter time T ≤ preset threshold Q1, it indicates that the potential risk level is low, and a level 2 warning report is generated.

2. According to the big data-based smart community video analysis system of claim 1, it is characterized by: The calculation process of the relative speed V' also includes the following process: Continuously obtain the average speeds Vr1, Vr2, Vr3, ..., Vrx of multiple unit time periods within a period of time when the target object enters the circular area Y; Using the formula Calculate the average speed of multiple unit time periods Vr1, Vr2, Vr3, ..., Vrx ; Randomly select the average speed of n time periods in multiple unit time periods, n>0, and calculate the deviation degree P of the speed in each time period. The calculation process is as follows: ; When P ≤ the preset threshold W, it means that the moving speed of the target object is balanced, so let ; When P>preset threshold W, it means that the moving speed of the target object fluctuates greatly, and the following processing is performed: Calculate the aggregation state β of the average speeds Vr1, Vr2, Vr3, ..., Vrx of multiple unit time periods. The specific calculation process is: ； make , obtain the average speed Vrx of β in this interval, and calculate the mean speed of the average speed Vrx of β in this interval , then let .

3. The smart community video analysis system based on big data according to claim 1 is characterized by: It also includes a database module, the database module is used to store historical data information of the movement of the target object, the historical data information includes the travel data information of the target object of the same type, and the establishment process of the circular area Y1 also includes: Obtain the driving data information of the target object, obtain the time t1 required for the target object to decelerate from the speed Vx to 0, and let the human body's reaction time be t2; Calculate the radius size R` of the circular area Y; When the target object is a single independent individual, the specific calculation process is: ; Let the preset size R = R`; When the target object is a plurality of independent entities, the real-time size data of the target object within a period of time is continuously obtained, which are L1, L2, L3, ..., Lx; Calculate the dimensional stability coefficient K of the target object. The specific calculation process is as follows: Using the formula Calculate the average size of the target object ; Calculate the stability coefficient K. The specific calculation process is: ； When the stability coefficient K≤the preset threshold U, the preset size R=R`; When the stability coefficient K> the preset threshold U, the preset size .

4. The smart community video analysis system based on big data according to claim 1 is characterized by: When the encounter time T is less than the preset threshold Q2, the data processing module further performs the following processing: Establish a virtual path, which is a straight line, and the two target objects are both within the virtual path, so that the two target objects move in the same or opposite directions, so that the positions of the two target objects in the width direction of the virtual path are consistent with the positions of the target objects on the actual path, and the path length dimension between the two target objects is Dx; Establish two circumscribed circles of the target object, namely circumscribed circle Y2 and circumscribed circle Y3, and obtain the radius size of circumscribed circle Y2 as Dr1 and the radius size of circumscribed circle Y3 as Dr2; Establish straight lines Ln1 and Ln2, both of which are tangent to circumscribed circles Y2 and Y3, and intersect at point B; Calculate the maximum distance between two target objects when they collide , the specific calculation process is: ； Wherein, α is the angle between any straight line Ln1 and straight line Ln2 and the geometric center line of the two target objects when the straight line is parallel to the straight line where the target object moves; Obtain the positional relationship between the two target objects on the actual path, transform the two target objects into the virtual path, and obtain the center distance L' of the two target objects; when When the potential risk level is changed, the first-level warning report is changed to a second-level warning report.

5. The smart community video analysis system based on big data according to claim 4 is characterized by: The process of establishing the circumscribed circle Y2 and the circumscribed circle Y3 also includes the following processing: Import standard path information of target object's moving trajectory; In the process of establishing the circumscribed circle Y2, the increment J1 of the moving path of the target object in the direction perpendicular to its moving direction is obtained, and the radius size of the circumscribed circle Y2 is changed to Dr3, and Dr3=Dr1+J1; During the process of establishing the circumscribed circle Y3, the increment J2 of the target object's moving path perpendicular to its moving direction is obtained, and the radius of the circumscribed circle Y3 is changed to Dr4, and Dr4=Dr1+J2.

6. The smart community video analysis system based on big data according to claim 1, characterized in that: When the preset threshold Q2≤encounter time T≤preset threshold Q1, the data processing module further performs the following processing: Importing real-time size data of two target objects respectively, wherein the real-time size data includes the length Li, width Di, and height Hi of the target objects; Use the formula Vix=Li·Di·Hi to calculate the volumes of the two target objects, which are Vi1 and Vi2 respectively; Calculate the volume difference S of the two target objects. The specific calculation process is: ； Import the real-time speeds of two target objects, which are V1 and V2 respectively; Calculate the speed difference G of the two target objects. The specific calculation process is: ； Assign a correction value e1 to the volume difference S, and a correction value e2 to the speed difference G, and e1+e2=1, e2>e1, and estimate the risk factor f when the target object collides. The specific estimation process is: ; When the risk factor f ≥ the preset threshold C, the potential risk level is changed and the second-level warning report is changed to a first-level warning report.

7. The smart community video analysis system based on big data according to claim 6 is characterized by: The calculation process of the volume of the target object also includes the following processing: For the same target object, its length Li, width Di, and height Hi are obtained at multiple points respectively, and the multiple points are evenly arranged, and the average value of the length Li, the average value of the width Di, and the average value of the height Hi are calculated respectively. When calculating the volume Vi, the average value of the length Li, the average value of the width Di, and the average value of the height Hi are taken respectively.

Images