Crowd supervision and evacuation guide robot based on video information

Abstract

The invention belongs to the field of public safety management research, and relates to a crowd supervision and evacuation guide robot based on video information, which comprises a robot body, and a data acquisition module, an obstacle avoidance module, a moving module, a smoke temperature sensing module, a data analysis processing module, a voice broadcast module and a wireless module which are carried on the robot body. According to the invention, global conditions can be mastered and intelligent crowd management can be carried out based on video information in public places, crowd conditions can be monitored at ordinary time, possible dangers can be judged, and early warning and interference can be carried out; when an accident occurs, the accident source and the crowd condition are known, the accident evolution condition and the utilization rate of each emergency exit are dynamically judged in real time, the crowd is guided to be evacuated to the proper emergency exit on the optimal evacuation path through acousto-optic pacification, and support is provided for safe and efficient evacuation of the crowd.

Description

Technical Field

[0001] The invention belongs to the research field of public safety management and relates to a crowd supervision and evacuation guiding robot based on video information. Background Art

[0002] Urban public places are an important part of the lives and work of urban residents, hosting a variety of important activities such as commerce, culture, entertainment, and sports. Urban public places such as large business districts (commercial streets and commercial complexes), transportation hubs, and large event venues have the characteristics of accommodating a large number of people, high mobility, limited space, and concentrated wealth. Large-scale, high-density crowd gatherings hide huge safety hazards. Crowd stampede accidents caused by excessive crowds will not only cause great casualties and serious property losses, but also bring adverse social impacts, which far exceed the impact of the accident itself.

[0003] Crowd safety management in crowded places has become a hot spot, focus and difficulty in public management. In order to ensure the life safety of large-scale crowds, it is urgent to start from daily management and emergency management. In daily management, it is necessary to achieve early detection of crowd safety hazards, assist managers in early decision-making, and avoid accidents; in emergency management, it is necessary to achieve safe and efficient crowd guidance when accidents occur, alleviate crowd panic, avoid crowded stampede accidents, and improve the initiative to respond to disasters.

[0004] In the existing technical solutions, video is an important means of crowd management. However, the traditional video surveillance system only provides simple functions such as video capture, storage and playback. There is a certain lag between the video information processing results and the on-site situation, and the monitoring cameras are fixedly arranged, resulting in monitoring blind spots in some areas, which cannot meet the functional requirements of real-time monitoring and early warning of dense crowds in public places and safe evacuation guidance. In addition, arranging multiple managers on site to supervise large crowds is also a common method, but managers are easily tired and cannot persist in long-term and efficient work. In addition, when faced with emergencies, they lack control over the overall information and find it difficult to make the best decision in the first time. Inadequate handling may occur. For example, when an accident occurs, inadequate evacuation guidance will lead to a reduced utilization rate of safety exits.

[0005] In summary, the main shortcomings of existing methods include:

[0006] In daily and emergency evacuation scenarios, video information does not play a sufficient role in guiding overall decision-making in crowd management; in daily scenarios, it is difficult to achieve all-weather, all-around supervision of dense crowds in public places; in evacuation scenarios, it is difficult to effectively calm and guide the crowd and improve evacuation efficiency.

[0007] Therefore, in order to strengthen the monitoring of the global crowd status in public places and solve the problem of effective evacuation of people when accidents occur, it is necessary to innovate a crowd supervision and evacuation guidance robot based on video information. Summary of the invention

[0008] The technical problem to be solved by the present invention is to provide an intelligent crowd management robot that can grasp the overall situation based on video information in public places, monitor the crowd situation in normal times, make judgments on possible dangers and issue early warnings and interventions; when an accident occurs, understand the source of the accident and the crowd situation, dynamically judge the evolution of the accident and the utilization rate of each safety exit in real time, and use sound and light to guide the crowd to evacuate to the appropriate safety exit on the optimal evacuation path, so as to provide support for the safe and efficient evacuation of the crowd.

[0009] The technical solution adopted by the present invention is as follows:

[0010] A crowd monitoring and evacuation guidance robot based on video information, comprising a robot body and a data acquisition module, an obstacle avoidance module, a movement module, a smoke temperature sensing module, a data analysis and processing module, a voice broadcast module and a wireless module mounted on the robot body;

[0011] The obstacle avoidance module is used to guide the robot to avoid obstacles and people;

[0012] The mobile module is used to control the movement of the robot according to the instructions of the obstacle avoidance module and the data processing and analysis module;

[0013] The smoke temperature sensing module is used to sense the gas composition and ambient temperature in the air in public places;

[0014] The data acquisition module is used to collect visual information of people in the field of view and the surrounding environment in real time;

[0015] The data processing and analysis module is used to analyze the collected video and sensor data, identify emergency situations and formulate evacuation plans;

[0016] The voice broadcast module is used to broadcast warning voice or evacuation guidance voice;

[0017] The wireless module is used to communicate wirelessly with the monitoring center to achieve real-time transmission and reception of data.

[0018] Furthermore, the data acquisition module includes at least one camera to acquire visual information.

[0019] Furthermore, the data processing and analysis module includes:

[0020] A storage unit, used to store video information received from the data acquisition module and video information transmitted through the wireless module;

[0021] Algorithmic cores that analyze stored video information to identify abnormal behavior or emergency situations;

[0022] The signal output unit is used to send a signal when the algorithm kernel determines that there is an unexpected event or accident.

[0023] Furthermore, the obstacle avoidance module includes at least one set of infrared sensors to obtain distance information; a built-in obstacle avoidance algorithm, specifically:

[0024] The distance between the robot body and the obstacle or the crowd obtained by the infrared sensor is compared with the preset first distance or second distance. When the distance to the obstacle is less than or equal to the first distance, an obstacle avoidance command is output; when the distance to the crowd is less than or equal to the second distance, a stay away command is output.

[0025] Furthermore, after receiving the voice broadcast signal from the data processing and analysis module, the voice broadcast module broadcasts a soothing and gentle voice.

[0026] Furthermore, the flue gas temperature sensing module includes a gas sensor and a temperature sensor, which issues a warning and sends a signal to a monitoring center when abnormal gas or abnormal temperature is sensed.

[0027] Furthermore, during the evacuation process, the smoke temperature sensor can sense the changes in gas and temperature, and guide the robot to guide the crowd to evacuate on a safe path.

[0028] Furthermore, the moving module includes a motor and a pair of symmetrical pulleys. The motor is placed inside the robot body to drive the pulleys and control the movement of the robot.

[0029] Furthermore, the wireless module is connected to a monitoring center related to the public place, obtains the video images of the monitoring cameras in the public place in real time and transmits them to the data processing and analysis module, and at the same time transmits the analysis results of the data processing and analysis module back to the monitoring center.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] With regard to the daily status supervision of the crowd, the present invention overcomes the difficulty of being unable to obtain the global and real-time crowd situation. By combining the monitoring video images and the images obtained by the robot camera, the crowd status can be monitored in all directions around the clock without blind spots, and timely warnings and responses can be made to emergencies and dangerous behaviors.

[0032] In view of the unknown time and space of the accident, the present invention overcomes the problem of being unable to accurately obtain the location and time of the accident. Through algorithmic analysis of monitoring video information and camera images, as well as sensing of gas, temperature and other indicators, the source of the accident and the evolution status of the accident can be analyzed more accurately, which is convenient for formulating evacuation guidance strategies.

[0033] With regard to the problem of crowd evacuation, the present invention overcomes the problem that the existing evacuation methods are less effective when an accident occurs. The analysis of video images and the perception of multiple sensors can determine the utilization of each safety exit, dangerous areas, crowd congestion points, etc., and reasonably plan the robot's movement path, so as to provide more correct guidance for the evacuation of the crowd, avoid the crowd from entering high-risk areas, improve the utilization rate of safety exits, and use appropriate sound and light for guidance to alleviate the panic of the crowd and reduce the occurrence of stampede accidents.

[0034] Regarding the information processing problem, the present invention overcomes the lag between video information processing and on-site conditions in public places. The robot can use its own algorithm to quickly and accurately process video information, determine the source of the accident and the crowd situation, and facilitate more efficient decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a diagram of a crowd monitoring and evacuation guiding robot based on video information according to an embodiment of the present invention.

[0036] Figure 2 This is a structural diagram of OpenPose in an embodiment of the present invention.

[0037] Figure 3 This is a diagram of the lightweight OpenPose network structure in an embodiment of the present invention.

[0038] In the figure, 1 is the robot body, 2 is the data acquisition module, 3 is the obstacle avoidance module, 4 is the smoke temperature sensing module, 5 is the voice broadcast module, 6 is the mobile module, 7 is the data processing and analysis module, and 8 is the wireless module. DETAILED DESCRIPTION

[0039] The principles and features of the present invention are described below in conjunction with the accompanying drawings and specific embodiments. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0040] Example 1

[0041] like Figure 1As shown, a crowd supervision and evacuation guidance robot based on video information includes a robot body 1 and a data acquisition module 2, an obstacle avoidance module 3, a smoke temperature sensing module 4, a voice broadcast module 5, a mobile module 6, a data analysis and processing module 7 and a wireless module 8 mounted on the robot body 1.

[0042] The data acquisition module 2 is located outside the robot body and is connected to the data analysis and processing module. It is used to collect visual information of people in the field of view and the surrounding environment in real time; it includes at least one camera 2, which is used to obtain the situation of people in the field of view and the surrounding environment.

[0043] The obstacle avoidance module 3 is located outside the robot body, connected to the data acquisition module and the mobile module, and is used to guide the robot to avoid obstacles and people; it includes at least one set of infrared sensors, which are used to obtain the distance information between the robot body and the place where the accident occurred or the crowd or obstacle. The obstacle avoidance module has a built-in obstacle avoidance algorithm, which is specifically: according to the distance between the robot body and the obstacle or the crowd obtained by the infrared sensor carried by the obstacle avoidance module, it is compared with the preset first distance or second distance, and an obstacle avoidance instruction is output when the distance to the obstacle is less than or equal to the first distance, and a distance instruction is output when the distance to the crowd is less than or equal to the second distance.

[0044] The wireless module 8 is located inside the robot body, connected to the data processing and analysis module, connected to the monitoring center related to the public place, and obtains the video images of all monitoring cameras in the public place in real time and transmits them to the data processing and analysis module. At the same time, it receives the analysis results of the data processing and analysis module and transmits them back to the monitoring center.

[0045] The mobile module 6 is connected to the obstacle avoidance module and the data processing and analysis module, and is used to control the movement of the robot according to the instructions of the obstacle avoidance module and the data processing and analysis module; the mobile module includes a motor and a pair of symmetrical pulleys, the motor is placed inside the robot body, and the symmetrical pulleys are arranged below the robot body. The pulleys are driven by the motor to control the movement of the robot.

[0046] The smoke temperature sensing module 4 is located outside the robot body, connected to the wireless module and the data analysis and processing module, and is used to sense the gas composition and ambient temperature in the air in public places; the smoke temperature sensing module includes a gas sensor and a temperature sensor, which are used to sense the gas composition and ambient temperature in the air in public places, and issue a warning when abnormal gas or abnormal temperature is sensed and send a signal to the command center through the wireless module, and issue an alarm after the command center confirms it. In addition, during the evacuation process, the changes in gas and temperature are sensed to guide the robot to guide the crowd to evacuate on a safe path.

[0047] The data processing and analysis module is located inside the robot body and is connected to the data acquisition module, the voice broadcast module, the wireless module, the smoke temperature sensing module, and the mobile module. It is used to analyze the collected video and sensing data, identify emergency situations, and formulate evacuation plans; the data processing and analysis module includes a storage module, an algorithm core, and a signal output unit. The storage unit is used to store video information received from the data acquisition module and the wireless module; the algorithm core is used to analyze the stored video information, and send the processing results from the wireless module to the monitoring center for retention; the signal output unit is used to send a signal when the algorithm core determines that there is an unexpected event or accident, so as to control the movement of the robot and broadcast the corresponding voice.

[0048] The main functions of the algorithm core video analysis in the present invention are crowd abnormal behavior detection and crowd density detection, and any algorithm in the prior art that can achieve this function can be used. A specific example is provided below:

[0049] The abnormal behavior detection algorithm is a pedestrian abnormal behavior recognition algorithm based on the open source human posture estimation framework (OpenPose):

[0050] OpenPose is a bottom-up two-dimensional human pose estimation algorithm proposed based on convolutional neural networks. The algorithm combines supervised learning methods to determine the joints of people in the image, which are key points, and then connects them to generate skeletal key points, thereby obtaining the corresponding connection graph.

[0051] The network structure diagram of OpenPose is as follows Figure 2 shown.

[0052] The network model uses the first ten layers of the VGG-19 deep neural network to initialize the extracted original feature map F of the image. The processed feature map F is then input into the network, and then the network is divided into two branches for input. One branch uses the convolutional neural network (CNN) to predict the heat map of the human body posture joints; the other branch uses CNN to obtain the partial affinity domain of all connected joints. The total number of stages in the network structure is T P , the total number of confidence map stages is T C .

[0053] In branch 1, each layer of the network plays an important role in estimating the key point confidence map. The confidence map is continuously detected starting from the latest partial affinity domain. The detection process is shown in formulas (1) and (2).

[0054]

[0055] Among them, ρ t For the convolutional neural network that performs inference in the tth iteration, perform TP After rounds of iterations, we get the confidence map S t To conduct T P After the first round of iteration, the confidence map obtained by repeated detection is obtained starting from the latest partial affinity domain.

[0056] In branch 2, each layer of the network is mainly used to predict the affinity domain of the key point part, as shown in formula (3).

[0057]

[0058] In the formula, Represents the convolutional neural network for reasoning at layer t. In each subsequent layer, the prediction results from the previous convolutional network are combined with the feature map F of the original image to accurately predict L t .

[0059] The multi-stage convolutional network is repeatedly iterated until the network converges. When it iterates to the last layer, the confidence map and partial affinity field are obtained, that is, the position information of all key points in the image and the direction vectors of adjacent key points can be obtained. The partial affinity field (PAF) is a two-dimensional vector field, whose main function is to describe the position of the limbs and their current direction. In the subsequent loop, the input is the output of the previous two-way branch loop and the original feature map F, and then the Hungarian algorithm is used for further calculation and analysis to output the two-dimensional coordinates of the human skeleton key points and limbs. Next, it is necessary to determine whether the obtained key points belong to the same person. According to the key points, the continuous pixel points d j1 and d j2 The PAF between them is used to make the judgment, and the judgment formula is shown in formula (4).

[0060]

[0061] Among them, E is the obtained PAF result, p(u) is the continuous pixel point d j1 and d j2 The pixels between are as shown in formula (5).

[0062] p(u)=(1-u)d j1 +ud j2 (5)

[0063] like Figure 3 As shown in Figure 2, a loss function is added at the end of each stage to help the convolutional network quickly and accurately estimate the partial affinity domains and confidence maps in the two branches in repeated iterations. The loss functions of the confidence map branch and partial affinity domain at stage t are shown in formulas (6) and (7).

[0064]

[0065] In the above formula, represents the prediction result of PAF in round t, It represents the position of the limbs. Represents the prediction result of the key point confidence map of round t, is the true key point confidence map, W(*) is the binary mask obtained when pixel p lacks annotation, and its main function is to minimize the negative impact caused by the accurate prediction of training true positives. represents the L2 loss function of PAF after t rounds of iterations, represents the L2 loss function of the confidence map after t rounds of iterations, and finally formula (8) can be derived.

[0066]

[0067] In order to evaluate f S Indicators need to generate S at the already marked key points during training * , i.e., confidence maps, where each confidence map is a two-dimensional map of the confidence of a particular body part in a given pixel. Assume represents the true position of the jth body part of the kth person in the image, then The value at p is defined as formula (9):

[0068]

[0069] Where σ is the expansion of the control peak, is the credibility of the jth body part of the kth person.

[0070] Then the credibility confidence map in the network prediction result can be represented by the set of each independent confidence map after passing the maximum operator, and the prediction result obtained is As shown in formula (10).

[0071]

[0072] Finally, the floating point operations per second (FLOPS) of the openpose network model are calculated, as shown in formula (11).

[0073] FLOPS=H×W×n×(2×K×K×c-1) (11)

[0074] Among them, K×K represents the size of the convolution kernel, H×W represents the size of the feature map, and c and n represent the number of input and output feature map channels respectively.

[0075] The pedestrian abnormal behavior recognition algorithm based on openpose posture estimation can analyze the behavior of the crowd

[0076] Features such as trajectory, posture, and joint points are used to identify abnormal behaviors that do not conform to normal behavior patterns, such as crowd gathering, running, pushing, and crowding.

[0077] The crowd density detection algorithm is the DSNet algorithm:

[0078] DSNet consists of a backbone network as a feature extractor, three dense dilated convolutional blocks stacked by dense residual connections to enlarge the scale diversity, and three convolutional layers for crowd density map regression.

[0079] The backbone network used is the first ten layers of VGG-16, and three pooling layers. A new dense dilated convolution block is used, which contains three dilated convolution layers with dilation rates of 1, 2, and 3. This setting can retain information from denser scales and has a smaller difference in receptive field size. Each dilated layer within the block is tightly connected to other layers, so each layer can access all subsequent layers and pass information that needs to be preserved. After dense connection, the scale diversity obtained increases. The architecture is improved by dense residual connections to further improve the information flow.

[0080] The multi-scale density level consistency loss is combined with the Euclidean loss to measure the global and local consistency. The Euclidean distance is used to measure the pixel-level estimation error between the estimated density map and the true value. The loss function L e The definition is shown in formula (12):

[0081]

[0082] Where N is the number of images in a batch, G(X i ; θ) is the training image X i The estimated density map of , with parameter θ. Yes X i The real density map.

[0083] In addition to the pixel-level loss function, the global and local density level consistency between the estimated density map and the true value is also considered. The newly proposed training loss L c The definition is shown in formula (13):

[0084]

[0085] Where S is the number of scale levels used for consistency checking, P ave is the average pooling operation, k j Specifies the output size for average pooling.

[0086] By weighted summing the above two loss functions, the entire network is trained using the following objective function L, as shown in formula (14):

[0087] L=L e +λL c (14)

[0088] where λ is a weight that balances the consistency loss between pixel and density levels.

[0089] For scene images with dense crowds in the dataset, geometric adaptive kernel processing is used to generate density maps, while for images with relatively sparse crowds in the dataset, fixed Gaussian kernel is used to generate density maps.

[0090] During testing, the entire image is input into the network to generate an estimated density map. The mean absolute error (MAE) and mean square error (MSE) are used to evaluate the network performance. MAE reflects the accuracy of the model, while MSE reflects the robustness of the model. The lower the value, the better the performance. The definitions of these two indicators are shown in equations (15) and (16):

[0091]

[0092] Where N is the number of images in the test set, C i represents the predicted count, and Indicates the actual count value.

[0093] The voice broadcast module is located inside the robot body and is connected to the data processing and analysis module. It includes a receiver and a playback device. After the receiver receives the signal transmitted by the data analysis and processing module, the playback device uses a soothing and soft voice (such as a 45-decibel female voice) to broadcast a warning voice (your behavior may cause danger, please stop immediately) or an evacuation guidance voice (an accident has occurred at location xx, please stay calm and don't panic, follow me to the xx safety exit for orderly evacuation) to alleviate the panic of the crowd and guide the crowd to evacuate.

[0094] Among the above parts, the camera is assembled on the top of the robot and connected to the data processing and analysis module of its chip core. The obstacle avoidance module is assembled on both sides of the robot. The robot chassis is equipped with a motor and a pulley connected to it to drive the robot. The wireless module is implanted in the robot core, and the robot body shell is assembled by welding with flame-retardant materials. The voice broadcast module is composed of a loudspeaker and a playback device and is set inside the robot. After receiving the internal command of the robot, it sends out the corresponding voice.

[0095] The robot's wireless module signal is connected to the monitoring center in the public place in advance to obtain the real-time video images of each monitoring camera. At the same time, the robot's external infrared camera obtains the images of the people and environment observed in its field of view.

[0096] In daily situations, a fixed cruise route in a public place is input into the robot, so that it patrols in the public place according to the preset path. The camera carried by the robot obtains the video images on its route and the images of each surveillance camera at the same time. The obtained images are analyzed by the preset computer vision algorithm of the data processing and analysis module to determine the current behavior status of the crowd. If the crowd is judged to have dangerous actions or abnormal status after image analysis, the robot will automatically cruise to the abnormal location, and the voice broadcast module will broadcast the voice to warn the crowd, and transmit the analysis and processing results back to the command center.

[0097] In the event of an accident, the robot switches to emergency mode. The robot obtains the monitoring images in the station through the wireless module and analyzes the images through the data analysis and processing module. At the same time, it combines the toxic gas and temperature sensing modules to determine the approximate location of the accident and transmits the processing results to the command center through the wireless module. The robot processes the images through the preset computer vision algorithm to form a density map of the crowd distribution near each safety exit in the public place, and determines the safety exits with low utilization and safe evacuation paths. The robot plays a voice to comfort the crowd through the voice broadcast module, and guides the crowd to evacuate to the appropriate safety exit on the optimal path.

[0098] In addition, the functions of each module of the present invention are not limited to a specific algorithm implementation, but can flexibly adopt including but not limited to traditional machine learning algorithms, deep learning algorithms, and any other algorithms that can meet the performance requirements of the present invention. Through this design, the present invention can adapt to different application scenarios and requirements, providing a high degree of flexibility and scalability.

[0099] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A crowd monitoring and evacuation guidance robot based on video information, characterized in that: It includes a robot body and a data acquisition module, an obstacle avoidance module, a mobile module, a smoke temperature sensing module, a data analysis and processing module, a voice broadcast module and a wireless module mounted on the robot body; The obstacle avoidance module is used to guide the robot to avoid obstacles and people; The mobile module is used to control the movement of the robot according to the instructions of the obstacle avoidance module and the data processing and analysis module; The smoke temperature sensing module is used to sense the gas composition and ambient temperature in the air in public places; The data acquisition module is used to collect visual information of people in the field of view and the surrounding environment in real time; The data processing and analysis module is used to analyze the collected video and sensor data, identify emergency situations and formulate evacuation plans; The voice broadcast module is used to broadcast warning voice or evacuation guidance voice; The wireless module is used to communicate wirelessly with a public place monitoring center to achieve real-time transmission and reception of data.

2. The crowd supervision and evacuation guidance robot according to claim 1, characterized in that: The data acquisition module includes at least one camera to acquire visual information.

3. The crowd supervision and evacuation guidance robot according to claim 1, characterized in that: The data processing and analysis module includes: A storage unit, used to store video information received from the data acquisition module and video information transmitted through the wireless module; Algorithmic cores that analyze stored video information to identify abnormal behavior or emergency situations; The signal output unit is used to send a signal when the algorithm kernel determines that there is an unexpected event or accident.

4. The crowd supervision and evacuation guidance robot according to claim 1, characterized in that: The obstacle avoidance module includes at least one set of infrared sensors to obtain distance information; a built-in obstacle avoidance algorithm, specifically: The distance between the robot body and the obstacle or the crowd obtained by the infrared sensor is compared with the preset first distance or second distance. When the distance to the obstacle is less than or equal to the first distance, an obstacle avoidance command is output; when the distance to the crowd is less than or equal to the second distance, a stay away command is output.

5. The crowd supervision and evacuation guidance robot according to claim 1, characterized in that: After receiving the voice broadcast signal from the data processing and analysis module, the voice broadcast module broadcasts the voice.

6. The crowd supervision and evacuation guidance robot according to claim 1, characterized in that: The flue gas temperature sensing module includes a gas sensor and a temperature sensor, and issues a warning and sends a signal to a monitoring center when abnormal gas or abnormal temperature is sensed.

7. The crowd supervision and evacuation guidance robot according to claim 1, characterized in that: During the evacuation process, the smoke temperature sensor can sense the changes in gas and temperature, and guide the robot to guide the crowd to evacuate on a safe path.

8. The crowd supervision and evacuation guidance robot according to claim 1, characterized in that: The moving module includes a motor and a pair of symmetrical pulleys. The motor is placed inside the robot body to drive the pulleys and control the movement of the robot.

9. The crowd supervision and evacuation guidance robot according to claim 1, characterized in that: The wireless module is connected to a monitoring center related to the public place, obtains the video images of the monitoring cameras in the public place in real time and transmits them to the data processing and analysis module, and at the same time transmits the analysis results of the data processing and analysis module back to the monitoring center.

Images