Method and device for monitoring falling objects from high altitude for smart city management

Abstract

The present invention discloses a method and device for monitoring falling objects from a high altitude for smart city management, and relates to the field of communication technology. The method comprises: reading the response accuracy of the monitoring system, and configuring joint sensors such as vision, infrared and radar according to different intervals of the building. Using environmental sensors to monitor the environmental conditions of the sensor location and initialize the acquisition. When the sensor is triggered, a dynamic tracking framework is established, and the focal length and acquisition frequency are automatically adjusted for monitoring to obtain the first batch of monitoring data. The sensor scans the lower area and performs joint monitoring in combination with the first batch of data to obtain the second batch of data. A trajectory diagram is drawn based on the two batches of data, and a falling object trajectory image is generated through a trajectory backtracking algorithm to perform alarm management of falling objects from a high altitude. The technical problem that the existing high-altitude falling object monitoring system cannot accurately and real-time monitor and effectively track falling objects is solved, and the technical effect of accurately monitoring the trajectory of falling objects and timely alarming is achieved, thereby improving the safety assurance capability in urban management.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a method and device for monitoring falling objects from a high altitude for smart city management. Background Art

[0002] In the management of modern smart cities, the safety issue of falling objects from high altitudes has received increasing attention, especially in areas with dense high-rise buildings. Traditional monitoring methods mainly rely on fixed cameras or personnel patrols, which are relatively lagging in real-time monitoring, positioning and trajectory tracking of falling objects from high altitudes, making it difficult to provide early warnings and handle them in a timely and effective manner. The existing monitoring system has low monitoring accuracy and is unable to dynamically capture the movement trajectory of falling objects from high altitudes. In addition, the system's response effect is poor when faced with different environmental conditions such as light changes or complex climates. In addition, the system's sensor configuration and data processing capabilities are limited, and it is impossible to quickly fuse and analyze multi-dimensional information in real time, resulting in low monitoring efficiency. In order to solve these problems, there is an urgent need for a technical solution that can improve the accuracy and response speed of falling object monitoring in order to enhance the overall capabilities of urban safety management.

[0003] In the current relevant technologies, there is a technical problem that the high-altitude falling object monitoring system cannot accurately and real-time monitor and effectively track falling objects. Summary of the invention

[0004] This application provides a method and device for monitoring falling objects from a high altitude for smart city management, adopts the response accuracy of the reading monitoring system, and configures joint sensors such as vision, infrared and radar according to different intervals of the building. Use environmental sensors to monitor the environmental conditions of the sensor location and initialize the acquisition. When the sensor is triggered, a dynamic tracking framework is established, and the focal length and acquisition frequency are automatically adjusted for monitoring to obtain the first batch of monitoring data. Subsequently, the sensor scans the lower area and performs joint monitoring in combination with the first batch of data to obtain the second batch of data. Based on these two batches of data, a trajectory map is drawn, and a falling object trajectory image is generated through a trajectory backtracking algorithm to perform alarm management of falling objects from a high altitude, thereby achieving the technical effect of accurately monitoring the trajectory of falling objects and timely alarming, and improving the safety assurance capabilities in urban management.

[0005] This application provides a high-altitude falling object monitoring method for smart city management, including:

[0006] Read the response accuracy of the high-altitude falling object monitoring, divide the target building into intervals based on the response accuracy, and configure the joint sensor using the interval division result and the response accuracy. The joint sensor includes a visual sensor, an infrared sensor, and a radar sensor; use the environmental sensor to perform environmental monitoring of the joint sensor position, synchronize the environmental monitoring result to the joint sensor, and perform acquisition initialization of the joint sensor; when the joint sensor is triggered, establish a dynamic tracking frame of the trigger object, and configure the automatic tracking focal length and acquisition frequency with the dynamic tracking frame, perform tracking acquisition, and establish a first acquisition result; based on the interval division result, put the joint sensor of the lower interval into a scanning state, configure the scanning focus point using the first acquisition result, perform joint monitoring of the lower interval, and establish a second acquisition result; draw a trajectory based on the first acquisition result and the second acquisition result, and use the trajectory backtracking network to perform trajectory backtracking fitting, establish a trajectory image of the trigger object according to the backtracking fitting result and the trajectory drawing result, and perform high-altitude falling object alarm management based on the trajectory image.

[0007] This application also provides a high-altitude falling object monitoring device for smart city management, including:

[0008] A joint sensor configuration module, the joint sensor configuration module is used to read the response accuracy of high-altitude falling object monitoring, divide the target building into intervals based on the response accuracy, and configure the joint sensor using the interval division result and the response accuracy. The joint sensor includes a visual sensor, an infrared sensor, and a radar sensor; an acquisition initialization execution module, the acquisition initialization execution module is used to use the environmental sensor to perform environmental monitoring of the joint sensor position, synchronize the environmental monitoring results to the joint sensor, and perform acquisition initialization of the joint sensor; a dynamic tracking frame establishment module, the dynamic tracking frame establishment module is used to establish a dynamic tracking frame of the trigger object when the joint sensor is triggered, The dynamic tracking frame is used to configure the automatic tracking focal length and acquisition frequency, perform tracking acquisition, and establish a first acquisition result; a joint monitoring execution module is used to place the joint sensor of the lower interval in a scanning state based on the interval division result, configure the scanning focus point using the first acquisition result, and then perform joint monitoring of the lower interval to establish a second acquisition result; a trajectory backtracking fitting module is used to draw the trajectory based on the first acquisition result and the second acquisition result, and perform trajectory backtracking fitting using the trajectory backtracking network, establish a trajectory image of the trigger object according to the backtracking fitting result and the trajectory drawing result, and perform high-altitude falling object alarm management based on the trajectory image.

[0009] The method and device for monitoring falling objects from high altitude for smart city management proposed in this application first reads the response accuracy of the monitoring system, and configures joint sensors such as vision, infrared, and radar according to different intervals of the building. Use environmental sensors to monitor the environmental conditions of the sensor location and initialize the acquisition. When the sensor is triggered, a dynamic tracking framework is established, and the focal length and acquisition frequency are automatically adjusted for monitoring to obtain the first batch of monitoring data. Subsequently, the sensor scans the lower area and performs joint monitoring in combination with the first batch of data to obtain the second batch of data. Based on these two batches of data, a trajectory map is drawn, and a falling object trajectory image is generated through a trajectory backtracing algorithm to perform alarm management of falling objects from high altitude, achieving the technical effect of accurately monitoring the trajectory of falling objects and timely alarming, and improving the safety assurance capabilities in urban management. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the device according to the embodiment of the present application. It should be understood that the previous or following operations are not necessarily performed precisely in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0011] Figure 1 A flowchart of a method for monitoring falling objects from a high altitude for smart city management provided in an embodiment of the present application;

[0012] Figure 2 A schematic diagram of the structure of a high-altitude falling object monitoring device for smart city management provided in an embodiment of the present application.

[0013] Description of reference numerals: joint sensor configuration module 10 , acquisition initialization execution module 20 , dynamic tracking frame establishment module 30 , joint monitoring execution module 40 , trajectory backtracking fitting module 50 . DETAILED DESCRIPTION

[0014] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0015] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0016] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, device, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0017] The present application embodiment provides a method for monitoring falling objects from a high altitude for smart city management, such as Figure 1 As shown, the method includes:

[0018] Step S100, read the response accuracy of the high-altitude falling object monitoring, divide the target building into intervals based on the response accuracy, and use the interval division result and response accuracy to configure the joint sensor, which includes a visual sensor, an infrared sensor, and a radar sensor. Specifically, first read the response accuracy of the high-altitude falling object monitoring, analyze the impact on the target building according to the accuracy, and divide the intervals, such as by floor or facade structure. Then, consider the characteristics of visual, infrared, and radar sensors and monitoring requirements to determine their distribution in each interval. Visual sensors are suitable for areas with good lighting, infrared sensors are used at night or in low-light areas, and radar sensors are used for key areas of high-precision monitoring trajectories. At the same time, it is necessary to consider the collaborative work of sensors to improve the monitoring effect.

[0019] Step S200, using the environmental sensor to perform environmental monitoring at the location of the joint sensor, synchronizing the environmental monitoring results to the joint sensor, and performing acquisition initialization of the joint sensor. Specifically, the environmental sensor is used to determine and monitor in real time the temperature, humidity, light intensity, wind speed, wind direction and other environmental parameters at the location of the joint sensor, and the results are synchronized to the joint sensor via wired or wireless means. After the control unit of the joint sensor parses and analyzes the data, the visual sensor adjusts the sensitivity and lens settings according to the light intensity and temperature and humidity, the infrared sensor adjusts the thermal sensitivity and protective heat dissipation settings according to the temperature and temperature and humidity, wind speed and direction, and the radar sensor adjusts the radar wave transmission direction angle, power and receiving sensitivity according to the wind speed and direction and temperature and humidity.

[0020] Step S300, when the joint sensor is triggered, a dynamic tracking frame of the trigger object is established, and the dynamic tracking frame is used to configure the automatic tracking focal length and acquisition frequency, perform tracking acquisition, and establish the first acquisition result. Specifically, when the joint sensor is triggered, the acquisition initialization result is used to configure the response priority of the visual, infrared, and radar sensors and initialize the corresponding trigger threshold. The trigger threshold is used to fuse the real-time monitoring results to trigger identification, and the sensor with the highest response priority is called to establish a dynamic tracking frame. The dynamic tracking frame is used to collect the features of the trigger object to establish a feature data set and authenticate the fall. After obtaining the position coordinates, the relative distance with the visual sensor position is calculated to configure the automatic tracking focal length. At the same time, the visual sensor acquisition frequency is configured based on the feature data set and the ambient light status. The tracking acquisition is performed using the automatic tracking focal length and acquisition frequency. Other sensors also collect according to their own logic to jointly establish the first acquisition result.

[0021] In one possible implementation, when the joint sensor is triggered, a dynamic tracking frame of the trigger object is established, and the automatic tracking focal length and acquisition frequency are configured with the dynamic tracking frame, tracking acquisition is performed, and a first acquisition result is established. Step S300 further includes step S310, using the acquisition initialization result to configure the response priority level of the visual sensor, infrared sensor, and radar sensor in the joint sensor, and initialize the trigger threshold based on the response priority level, and the trigger threshold includes the pixel change rate threshold, heat source displacement threshold, and speed change threshold of the visual sensor. Specifically, the collection initialization results are used to configure the response priority of the visual sensor, infrared sensor, and radar sensor in the joint sensor. The collection initialization results include relevant information obtained by the sensor during the environmental monitoring and initialization process, such as the impact of environmental light intensity, temperature, humidity, wind speed and direction on sensor performance. The response priority configuration of the visual sensor will take into account its dependence on light intensity and the current ambient light conditions. If the ambient light is sufficient and suitable for the visual sensor to work, it may be given a higher priority; if the light is insufficient, the priority may be lowered. The response priority of the infrared sensor will be based on factors such as ambient temperature and whether there is a thermal interference source. In an environment with suitable temperature and fewer thermal interference sources, its ability to detect heat sources is stronger and may have a higher priority. The response priority of the radar sensor will take into account the electromagnetic interference in the environment and the motion characteristics of the target object. If the electromagnetic interference is small and the target object moves quickly and requires high-precision position and speed detection, the radar sensor may be given priority. Initialize the trigger threshold based on the response priority of the visual sensor. For the visual sensor, set the pixel change rate threshold. The threshold is determined based on the resolution of the visual sensor and the accuracy requirements for trigger detection. For example, if the resolution of the visual sensor is high and it is required to be able to detect small changes in objects, the pixel change rate threshold may be set to a lower value, such as 10% (that is, when the pixel change in the image exceeds 10%, it is considered that there may be a trigger). For the infrared sensor, initialize the heat source displacement threshold. The determination of this threshold takes into account the sensitivity of the infrared sensor to heat source detection and the thermal radiation characteristics of the target object. If the infrared sensor has a high sensitivity to heat source detection and the thermal radiation of the target object is strong, the heat source displacement threshold may be set to a smaller value, such as 5 cm (that is, it is triggered when the detected heat source position moves more than 5 cm relative to the initial position). For the radar sensor, set the speed change threshold. The threshold is determined based on the accuracy of the radar sensor's object speed detection and the expected movement speed range of the target object. For example, if the radar sensor can accurately detect small speed changes and the target object may have a high movement speed, the speed change threshold may be set to 2 meters / second (that is, it is triggered when the detected object speed changes by more than 2 meters / second relative to the initial speed).

[0022] Step S320, using the trigger threshold to perform fusion trigger identification on the real-time monitoring results of the joint sensor, calling the sensor with the highest response priority based on the fusion trigger identification result, and establishing a dynamic tracking frame based on the sensor with the highest response priority and the fusion trigger identification result. Specifically, using the set trigger threshold to perform fusion trigger recognition on the real-time monitoring results of the joint sensor means that the pixel change of the visual sensor, the heat source displacement of the infrared sensor, and the speed change of the radar sensor must be considered at the same time. If the pixel change rate of the visual sensor exceeds the set threshold, or the heat source displacement of the infrared sensor exceeds its threshold, or the speed change of the radar sensor exceeds its threshold, this information is combined to determine whether a trigger object actually exists and its approximate position and motion state, and the sensor with the highest response priority is called based on the fusion trigger recognition result. For example, if it is found that the pixel change rate of the visual sensor exceeds the threshold first and its response priority is the highest during the fusion trigger recognition process, then the visual sensor is called for subsequent operations, and a dynamic tracking frame is established based on the called sensor with the highest response level and the fusion trigger recognition result. If the visual sensor is called, the dynamic tracking frame may be established based on the initial position and approximate outline of the trigger object captured by the visual sensor, and as the trigger object moves, the dynamic tracking frame will be dynamically adjusted according to the image information subsequently captured by the visual sensor, always surrounding the trigger object, so as to better track and collect relevant data.

[0023] In a possible implementation, when the joint sensor is triggered, a dynamic tracking frame of the trigger object is established, and the dynamic tracking frame is used to configure the automatic tracking focal length and acquisition frequency, perform tracking acquisition, and establish a first acquisition result. Step S300 further includes step S330, which performs feature acquisition of the trigger object based on the dynamic tracking frame to establish a feature data set of the trigger object. Specifically, the feature acquisition of the trigger object is performed based on the dynamic tracking frame, and the dynamic tracking frame surrounds the trigger object, providing a clear area range for feature acquisition. Within this range, the visual sensor can obtain the image information of the trigger object, including shape, color, texture, etc.; the infrared sensor can detect the thermal radiation characteristics of the trigger object; the radar sensor can obtain the position, speed and other motion characteristics of the trigger object, and integrate the various feature information about the trigger object collected from different sensors to establish a feature data set of the trigger object. The data set is the basis for subsequent analysis and processing. It contains multi-faceted feature information of the trigger object and can more comprehensively describe the state of the trigger object.

[0024] Step S340, after performing fall authentication of the trigger object with the feature data set, obtain the position coordinates of the dynamic tracking frame. Specifically, the fall authentication of the trigger object is performed with the feature data set, and the relevant feature information in the feature data set is analyzed to determine whether the trigger object is in a falling state. For example, observe whether the speed of the trigger object has an obvious increasing trend in the vertical direction, and whether its position is constantly decreasing, etc., and combine this information to determine whether the trigger object is falling. After completing the fall authentication, obtain the position coordinates of the dynamic tracking frame. The position coordinates of the dynamic tracking frame reflect the approximate position of the trigger object in space, which is very important for subsequent tracking and focus adjustment operations.

[0025] Step S350, relative distance calculation is performed using the position coordinates and the position of the visual sensor, and the automatic tracking focal length of the visual sensor is established based on the relative distance calculation result. Specifically, relative distance calculation is performed using the position coordinates and the position of the visual sensor, the position of the visual sensor is known, and the relative distance between the position coordinates of the dynamic tracking frame where the trigger object is located and the position of the visual sensor can be calculated. The automatic tracking focal length of the visual sensor is established based on the relative distance calculation result. If the relative distance is far, in order to clearly observe the trigger object, the focal length of the visual sensor needs to be adjusted to a long focus mode, that is, the automatic tracking focal length is set to a larger value; if the relative distance is close, the focal length is adjusted to a short focus mode, and the automatic tracking focal length is set to a smaller value.

[0026] Step S360, establish the visual sensor acquisition frequency based on the feature data set and the ambient light state, perform tracking acquisition using the automatic tracking focal length and the visual sensor acquisition frequency, establish the first visual acquisition result, and use the first visual acquisition result as a collection dimension data of the first acquisition result. Specifically, the visual sensor acquisition frequency is established based on the feature data set and the ambient light state, and the information such as the movement speed of the trigger object is obtained from the feature data set. If the movement speed of the trigger object is fast, in order to be able to completely capture the movement process of the trigger object, it may be necessary to increase the acquisition frequency; at the same time, consider the ambient light state. If the ambient light is dark, in order to obtain a clearer image, it may be necessary to reduce the acquisition frequency. These factors are combined to determine the appropriate acquisition frequency, and the automatic tracking focal length and the visual sensor acquisition frequency are used to perform tracking acquisition. The visual sensor tracks and shoots the trigger object according to the set automatic tracking focal length and acquisition frequency, obtains a series of image information, and establishes the first visual acquisition result. The first visual acquisition result contains images of the trigger object at different times during the tracking process. These images can reflect the movement process and state changes of the trigger object. The first visual acquisition result is used as a collection dimension data of the first acquisition result, and together with the data collected by other sensors, it constitutes the first acquisition result.

[0027] In a possible implementation, a visual sensor acquisition frequency is established based on a feature data set and an ambient light state, tracking acquisition is performed using an automatic tracking focal length and a visual sensor acquisition frequency, a first visual acquisition result is established, and the first visual acquisition result is used as an acquisition dimension data of the first acquisition result. Step S360 further includes step S361, obtaining the trigger object speed data in the feature data set, and using the trigger object speed data as the first decision data. Specifically, the trigger object speed data is obtained from the feature data set, and the feature data set contains various information about the trigger object collected from multiple sensors, among which the speed data is a key item. The speed data is obtained by analyzing and calculating the position change of the trigger object in the continuous image by a radar sensor or a visual sensor. The obtained trigger object speed data is used as the first decision data. The speed of the trigger object is very important for determining the acquisition frequency of the visual sensor, because if the speed of the trigger object is fast, it is necessary to collect images more frequently to accurately record its motion trajectory; if the speed is slow, the acquisition frequency can be appropriately reduced.

[0028] Step S362, obtain the trigger brightness and ambient light status in the feature data set to perform an environmental fusion evaluation of the trigger, and establish the second decision data based on the environmental fusion evaluation result. Specifically, the trigger brightness information is obtained from the feature data set, and the ambient light status of the current environment is obtained at the same time. The trigger brightness information can reflect the reflection characteristics of the trigger itself and its interaction with the ambient light. The ambient light status includes factors such as light intensity and light direction. The environmental fusion evaluation of the trigger is performed based on the trigger brightness and ambient light status. In low-light conditions, since the details of the object are difficult to identify, the system must not only consider reducing the acquisition frequency and increasing the exposure time to improve the image brightness, but also consider the impact of such adjustment on the overall fusion of the trigger and the environment. For example, if the exposure time is increased, the background may be too bright, which may reduce the contrast between the trigger and the background, affecting the recognition of the trigger. At this time, it is necessary to weigh the exposure. The increase in time. In high-dynamic scenes, if there are large light changes (such as shadows, strong light, etc.), the system captures more light changes by appropriately increasing the frame rate to ensure the continuity of the object's trajectory. At the same time, the system must also consider the changes in the visual features of the trigger under different light conditions. For example, strong light may cause the trigger to reflect, and shadows may cover some features of the trigger. It is necessary to evaluate the changes in the degree of integration of the trigger in the environment based on these situations, and establish the second decision data based on the environmental fusion evaluation results. The second decision data comprehensively considers factors such as the visibility of the trigger under the current ambient light conditions, the contrast with the environment, and the overall degree of integration, and is used to determine the visual sensor acquisition frequency together with the first decision data.

[0029] Step S363, establish the visual sensor acquisition frequency based on the first decision data and the second decision data. Specifically, the visual sensor acquisition frequency is established based on the first decision data (trigger speed data) and the second decision data (environmental fusion evaluation result). If the trigger speed is fast, the acquisition frequency tends to be increased according to the first decision data; but if the environmental fusion evaluation result shows that increasing the exposure time under low light conditions will seriously affect the trigger recognition, then it is necessary to balance between increasing the acquisition frequency and ensuring image quality. The acquisition frequency may be appropriately reduced to increase the exposure time, or other image processing technologies may be used to assist. If the environmental fusion evaluation result shows that it is in a high-dynamic scene and the light changes are complex, the frame rate tends to be increased according to the second decision data, but at the same time, it should be combined with the trigger speed. If the speed is slow, there is no need to excessively increase the frame rate to avoid data redundancy. By comprehensively considering various situations of the first decision data and the second decision data, a suitable visual sensor acquisition frequency is finally determined to ensure that while accurately recording the motion trajectory of the trigger, clear, recognizable image data that can reflect the relationship between the trigger and the environment can be obtained.

[0030] In a possible implementation, a visual sensor acquisition frequency is established based on a feature data set and an ambient light state, tracking acquisition is performed using an automatic tracking focal length and a visual sensor acquisition frequency, a first visual acquisition result is established, and the first visual acquisition result is used as an acquisition dimension data of the first acquisition result. Step S360 further includes step S364, establishing an infrared sensor acquisition frequency using the trigger object speed data and the ambient light state. Specifically, the infrared sensor acquisition frequency is established using the trigger object speed data and the ambient light state. The trigger object speed data is one of the key factors. If the trigger object is fast, a higher acquisition frequency is required to more accurately track its motion trajectory. The ambient light state also affects the infrared sensor acquisition frequency. In a low-light environment, the performance of the infrared sensor is relatively more stable. At this time, the acquisition frequency can be appropriately increased to obtain more thermal radiation information about the trigger object. In a strong light environment, although the infrared sensor is less affected by light, the strong light may produce certain thermal interference, and the acquisition frequency needs to be adjusted according to the specific situation. The frequency may need to be slightly reduced to avoid misjudgment.

[0031] Step S365, perform tracking acquisition based on the infrared sensor acquisition frequency, and establish a second infrared acquisition result. Specifically, based on the determined infrared sensor acquisition frequency, the tracking acquisition is performed, the infrared sensor continuously detects the thermal radiation signal of the trigger object at the set frequency, obtains the thermal characteristic information of the trigger object at different times, and organizes and records a series of thermal radiation data obtained during the tracking acquisition process to establish a second infrared acquisition result, which includes information such as the thermal radiation intensity and position (inferred by the position change of the thermal characteristic) of the trigger object at different time points, reflecting the change in the thermal characteristics and the motion state of the trigger object.

[0032] Step S366, call the environmental monitoring results, obtain radar signal interference, and establish the radar sensor acquisition frequency based on the radar signal interference and the trigger object speed data. Specifically, call the environmental monitoring results, obtain the radar signal interference, and establish the radar sensor acquisition frequency in combination with the trigger object speed data. Radar signal interference may come from reflections from other surrounding electronic devices, buildings and other objects. If the radar signal interference is strong, in order to ensure accurate detection of the position and speed of the trigger object, it is necessary to reduce the acquisition frequency to reduce the impact of the interference signal on the valid data; if the interference is weak and the trigger object is fast, the acquisition frequency can be appropriately increased to more accurately track the motion trajectory of the trigger object. The trigger object speed data also plays an important role. The faster the speed, the higher the acquisition frequency is required to capture its position changes in time.

[0033] Step S367, perform tracking acquisition based on the radar sensor acquisition frequency, and establish a third radar acquisition result. Specifically, based on the determined radar sensor acquisition frequency, perform tracking acquisition, the radar sensor transmits radar waves and receives reflected waves, and determines the position, speed and other parameters of the trigger object based on the time difference and frequency change of the reflected waves. The trigger object position, speed and other data obtained during the tracking acquisition process are integrated to establish the third radar acquisition result, which records in detail the position change sequence and speed change of the trigger object during the movement process, providing an important basis for subsequent trajectory analysis and processing.

[0034] Step S368, the second infrared acquisition result and the third radar acquisition result are used as two acquisition dimension data of the first acquisition result. Specifically, the second infrared acquisition result and the third radar acquisition result are used as two acquisition dimension data of the first acquisition result. The first acquisition result is a collection of multiple sensor data, wherein the infrared acquisition result provides the thermal feature information of the trigger object, and the radar acquisition result provides the position and speed information of the trigger object. Together with the visual sensor acquisition result (the first visual acquisition result, which may include the image features of the trigger object, etc.), they constitute a multi-dimensional description of the trigger object, providing comprehensive and accurate data support for subsequent trajectory drawing, analysis, and high-altitude falling object alarm management.

[0035] Step S400, based on the interval division result, the joint sensor of the lower interval is placed in a scanning state, and after the scanning focus point is configured using the first acquisition result, the joint monitoring of the lower interval is performed to establish a second acquisition result. Specifically, based on the interval division result, the scanning of the lower interval joint sensor is started, and then the scanning focus point is determined by analyzing the first acquisition result containing information such as the position and speed of the trigger object. The visual, infrared, and radar sensors of the lower interval work together accordingly. The visual sensor captures objects with appearance features, the infrared sensor detects objects with thermal radiation features, and the radar sensor determines the position and speed matching objects. Then, the image, thermal radiation, position and speed data obtained by each sensor are integrated, and finally a second acquisition result reflecting the situation of objects in the focus area of ​​the lower interval is established, which together with the first acquisition result constitutes a complete monitoring data chain to assist subsequent analysis.

[0036] Step S500, based on the first acquisition result and the second acquisition result, the trajectory is drawn, and the trajectory back-tracing network is used to perform trajectory back-tracing fitting, and the trajectory image of the trigger object is established according to the back-tracing fitting result and the trajectory drawing result, and the high-altitude falling object alarm management is performed based on the trajectory image. Specifically, based on the first acquisition result and the second acquisition result, the data is integrated and the trajectory of the trigger object is preliminarily drawn, and then the trajectory back-tracing network is used to call the interpolation channel to process the data, and then the back-tracing fitting is performed on the analysis layer according to various factors, and then the fitting result and the drawing result are merged to generate a trajectory image with a throwing point mark, and finally, based on this image analysis, it is judged whether the trajectory of the trigger object is dangerous, and if so, the alarm mechanism including sound and light alarms, push information, etc. is triggered to take countermeasures to avoid harm.

[0037] In one possible implementation, a trajectory is drawn based on the first acquisition result and the second acquisition result, and a trajectory back-tracing network is used to perform trajectory back-tracing fitting. A trajectory image of the trigger object is established according to the back-tracing fitting result and the trajectory drawing result. High-altitude falling object alarm management is performed based on the trajectory image. Step S500 further includes step S510, calling the interpolation channel of the trajectory back-tracing network, using the interpolation channel to interpolate the first acquisition result and the second acquisition result, and establishing an interpolation processing result. Specifically, the interpolation channel of the trajectory backtracking network is called. The trajectory backtracking network is a key part of the entire system for optimization and accurate trajectory analysis. Its interpolation channel is a functional module specifically used to deal with data missing and discontinuity problems. When a more accurate trajectory analysis of the acquisition results is required, this interpolation channel is first activated to prepare for subsequent data processing. The interpolation channel is used to interpolate the first acquisition result and the second acquisition result. The first acquisition result includes various data monitored by the trigger in the initial trigger area (such as the higher-level interval of a building), such as position, speed, and characteristics at different times; the second acquisition result is the data obtained by joint monitoring of the lower interval. Due to the frequency and accuracy limitations of sensor data collection, these data may be discontinuous or missing in the time series. In some situations, the interpolation channel will use mathematical algorithms (such as linear interpolation, polynomial interpolation, etc.) to estimate the value of missing data points based on the existing data points. For example, if the visual sensor fails to obtain the image data of the trigger object due to occlusion or other reasons in a certain period of time (resulting in missing position information), the interpolation channel can estimate the possible position of the trigger object during this period of time based on the position data of the previous and next moments and other sensors (such as the speed data of the radar sensor), so as to make the data more continuous and smooth in the time series. After interpolation processing, a more complete and continuous data set will be obtained, that is, the interpolation processing result. The result not only contains the valid data in the original acquisition result, but also supplements the missing data points, providing a richer and more reliable data basis for the subsequent precise trajectory analysis.

[0038] Step S520, synchronize the interpolation processing result to the analysis layer of the trajectory backtracking network, perform backtracking fitting, and establish the backtracking fitting result. Specifically, synchronize the interpolation processing result to the analysis layer of the trajectory backtracking network. The analysis layer is the part of the trajectory backtracking network that performs core calculations and model fitting. It requires accurate and complete data to execute complex algorithms. The interpolation processing result is transmitted from the interpolation channel to the analysis layer through the internal data transmission mechanism to ensure the timeliness and accuracy of the data. In the analysis layer, the backtracking fitting operation is performed. The analysis layer will use a series of advanced algorithms and models to process the interpolation processing results. These algorithms take into account a variety of factors, such as the physical characteristics of the trigger's movement (such as the laws of free fall under gravity), environmental factors (such as the impact of wind speed and wind direction on the trigger's movement trajectory), and the error characteristics of sensor data collection. By substituting the data in the interpolation processing results into these algorithms and models, the most likely true trajectory of the trigger at each moment is calculated. For example, for an object falling from a high building, the algorithm will optimize and correct the entire trajectory based on its initial velocity (obtained from the first collection result), gravitational acceleration, and possible air resistance (estimated in combination with environmental factors) to obtain a back-testing fitting result that is more in line with the actual situation. The result can more accurately reflect the true movement path of the trigger and reduce trajectory deviations caused by data errors and incompleteness.

[0039] Step S530, using the backtracking fitting result to generate a trajectory image, wherein the trajectory image is provided with a throwing point mark. Specifically, the trajectory image is generated using the backtracking fitting result, and the backtracking fitting result contains the precise position information of the trigger at each moment during the entire motion process, and the position information is plotted in a two-dimensional or three-dimensional coordinate system to form a continuous trajectory curve. For example, in the two-dimensional image, the plane position of the building is represented by the horizontal direction, and the height is represented by the vertical direction. A series of points from the initial trigger position to the possible landing position of the trigger are connected to intuitively display the motion trajectory of the trigger. According to the initial appearance position and motion state characteristics of the trigger in the first acquisition result, the position of the throwing point mark is determined. The mark is usually at the starting end of the trajectory image, and is represented by a special symbol (such as a circle, a triangle, etc.) or a color mark. The accurate display of the throwing point mark is very important for subsequent accident analysis and responsibility tracing. It can clarify the starting position of the trigger, which is helpful to determine the area or floor where the source of the high-altitude falling object incident is located. The trajectory image with the throwing point mark also provides clearer and more comprehensive information for relevant managers, so that they can quickly understand the general situation of the incident and take corresponding measures.

[0040] In a possible implementation, trajectory drawing is performed based on the first acquisition result and the second acquisition result, and trajectory backtracking fitting is performed using a trajectory backtracking network. A trajectory image of the trigger object is established based on the backtracking fitting result and the trajectory drawing result, and high-altitude falling object alarm management is performed based on the trajectory image. Step S500 further includes step S540, establishing digital communication with the built-in camera of the target building. Specifically, establishing digital communication with the built-in camera of the target building involves connecting the monitoring system to the camera pre-installed in the building through network technology (such as Ethernet, Wi-Fi, etc.). First, it is necessary to determine the network access method and communication protocol of the camera, and then perform corresponding configuration in the monitoring system to ensure that data can be transmitted stably and efficiently between the two. For example, if the camera supports Wi-Fi connection, the monitoring system needs to search and connect to the Wi-Fi network where the camera is located, and perform identity authentication and parameter settings to establish a reliable digital communication link.

[0041] Step S550, reading the real-time captured data of the built-in camera, performing a prediction analysis on the throwing tendency of the real-time captured data, and establishing a prediction analysis result. Specifically, after the communication is successfully established, the monitoring system can read the real-time captured data of the built-in camera. The camera continuously captures various areas in the building (such as corridors, balconies, etc. where falling objects may occur), and transmits these image data to the monitoring system in real time. After receiving these data, the monitoring system converts them into a format that can be analyzed and processed, such as decoding the video stream into a series of image frames, and performing throwing tendency prediction analysis on the real-time captured data. First, the target detection algorithm based on deep learning (such as YOLO series, FasterR-CNN, etc.) is used to identify objects and people in the image. The algorithm is pre-trained on a large amount of labeled data (including people, object images and labeling information in various scenes) to learn the characteristic patterns of different objects and people. For example, the YOLO algorithm divides the image into multiple grid units, each unit is responsible for predicting a certain number of object categories and their location information, and outputs the location and category probability of various objects in the image (such as children, windows, flower pots, etc.) through the forward propagation calculation of the network. After identifying the objects and people, the human posture estimation algorithm (such as OpenP ose, etc.) to analyze the posture of the character. The OpenPose algorithm builds a connection model of the human joints and calculates the position coordinates of each joint of the human body (such as the head, hands, feet, etc.) according to the contour, texture and other information of the character in the image, so as to determine the posture and action of the character. For example, by analyzing the position relationship of the joints, it can be judged whether the character is in an action posture related to high-altitude throwing, such as reaching, bending, approaching the edge, etc. Combining the results of target detection and human posture estimation, the behavior analysis algorithm is used to make a comprehensive judgment on the character's behavior. Based on the pre-set behavior pattern rules and machine learning models (such as hidden Markov models, support vector machines, etc.), the character's action sequence and the interaction relationship with the surrounding objects are analyzed. For example, if a child is detected approaching the edge of the balcony and the hand has a forward waving action posture, the behavior analysis algorithm will judge that it has a high tendency to throw according to the training model. According to the analysis results of the above series of algorithms, a prediction analysis result is established. The result not only includes a clear judgment on whether the person or object in the image has a tendency to throw (such as "throwing tendency" or "no throwing tendency"), but also includes relevant confidence information.The confidence level is calculated using the probability distribution learned by the algorithm model during training, and is used to indicate the reliability of the system's judgment results. For example, if the behavior analysis algorithm determines that a child has a high tendency to throw, the predictive analysis results will clearly mark this situation and give a higher confidence value (such as 85%). It will also record relevant image features (such as the child's appearance, environmental characteristics, etc.) and key data involved in the analysis process (such as human joint coordinates, object location information, etc.), providing detailed basis for subsequent early warning decisions and record storage.

[0042] Step S560, based on the prediction and analysis results, a voice reminder is given, and a capture record is generated. Specifically, a voice reminder is given based on the prediction and analysis results. When the prediction and analysis results show that there is a tendency to throw objects, the monitoring system immediately triggers the voice reminder function. The system pre-stores a series of voice prompt information (such as "Please note, do not throw objects from high places", "Dangerous behavior, please stop immediately", etc.), and selects appropriate voice files for playback according to different scenarios and judgment results. The voice reminder function is realized through a connected audio output device (such as a speaker), which converts the voice signal into sound and plays it in the building, promptly stops possible high-altitude throwing behaviors, reminds relevant personnel to pay attention to safety, and generates a capture record at the same time. The capture record covers a wealth of information, including the precise time of the throwing tendency event (accurate to milliseconds, obtained through the system clock), the specific location (determined according to the installation location of the camera and the geographic identification information in the image, such as the floor number, room number or district domain name, etc.), features of people or objects involved (obtained through image recognition and analysis results, such as a description of a child's appearance features, the shape and color of an object, etc.), and related images or video clips (directly extracting the part containing key information of the event from the real-time captured data and storing it in the original image format or compressed video format). The records will be stored in a local database (such as using relational database MySQL or non-relational database MongoDB to store structured and unstructured data) or a cloud server (uploading data over the network and using cloud storage services to ensure data reliability and scalability) for subsequent viewing, analysis, and use as evidence in possible investigations.

[0043] The embodiment of the present application adopts the response accuracy of the reading monitoring system, and configures joint sensors such as vision, infrared and radar according to different intervals of the building. Environmental sensors are used to monitor the environmental conditions of the sensor location and initialize the acquisition. When the sensor is triggered, a dynamic tracking framework is established, and the focal length and acquisition frequency are automatically adjusted for monitoring to obtain the first batch of monitoring data. Subsequently, the sensor scans the lower area and performs joint monitoring in combination with the first batch of data to obtain the second batch of data. Based on these two batches of data, a trajectory map is drawn, and a falling object trajectory image is generated through a trajectory backtracking algorithm to perform alarm management of falling objects from high altitudes, thereby achieving the technical effect of accurately monitoring the trajectory of falling objects and timely alarming, and improving the safety assurance capabilities in urban management.

[0044] In the above, refer to Figure 1 The high altitude falling object monitoring method for smart city management according to an embodiment of the present invention is described in detail. Figure 2 A high-altitude falling object monitoring device for smart city management according to an embodiment of the present invention is described.

[0045] The high-altitude falling object monitoring device for smart city management according to the embodiment of the present invention is used to solve the technical problem that the existing high-altitude falling object monitoring system cannot accurately and in real time monitor and effectively track falling objects, and achieves the technical effect of accurately monitoring the trajectory of falling objects and timely alarming, thereby improving the safety assurance capability in urban management. The high-altitude falling object monitoring device for smart city management includes: a joint sensor configuration module 10, an acquisition initialization execution module 20, a dynamic tracking frame establishment module 30, a joint monitoring execution module 40, and a trajectory backtracking fitting module 50.

[0046] The joint sensor configuration module 10 is used to read the response accuracy of high-altitude falling object monitoring, divide the target building into intervals based on the response accuracy, and configure the joint sensor using the interval division result and the response accuracy. The joint sensor includes a visual sensor, an infrared sensor, and a radar sensor.

[0047] The acquisition initialization execution module 20 is used to perform environmental monitoring of the joint sensor location using the environmental sensor, synchronize the environmental monitoring results to the joint sensor, and perform acquisition initialization of the joint sensor.

[0048] The dynamic tracking frame establishment module 30 is used to establish a dynamic tracking frame of the trigger object when the joint sensor is triggered, and configure the automatic tracking focal length and acquisition frequency with the dynamic tracking frame, perform tracking acquisition, and establish a first acquisition result.

[0049] The joint monitoring execution module 40 is used to place the joint sensor of the underlying interval in a scanning state based on the interval division result, configure the scanning focus point using the first acquisition result, perform joint monitoring of the underlying interval, and establish a second acquisition result.

[0050] The trajectory backtracking fitting module 50 is used to draw a trajectory based on the first acquisition result and the second acquisition result, and use the trajectory backtracking network to perform trajectory backtracking fitting, establish a trajectory image of the trigger object according to the backtracking fitting result and the trajectory drawing result, and perform high-altitude falling object alarm management based on the trajectory image.

[0051] The specific configuration of the dynamic tracking frame establishment module 30 will be described in detail below. As described above, when the joint sensor is triggered, a dynamic tracking frame of the trigger object is established, and the automatic tracking focal length and acquisition frequency are configured with the dynamic tracking frame, tracking acquisition is performed, and a first acquisition result is established. The dynamic tracking frame establishment module 30 further includes: a threshold trigger unit, the threshold trigger unit is used to configure the response priority level of the visual sensor, infrared sensor, and radar sensor in the joint sensor using the acquisition initialization result, and initialize the trigger threshold based on the response priority level, the trigger threshold includes the pixel change rate threshold, heat source displacement threshold, and speed change threshold of the visual sensor; a fusion trigger recognition unit, the fusion trigger recognition unit is used to perform fusion trigger recognition on the real-time monitoring results of the joint sensor using the trigger threshold, call the sensor with the highest response priority based on the fusion trigger recognition result, and establish a dynamic tracking frame based on the sensor with the highest response level and the fusion trigger recognition result.

[0052] Among them, the dynamic tracking frame establishment module 30 further includes: a feature data set establishment unit, the feature data set establishment unit is used to collect the features of the trigger object based on the dynamic tracking frame, and establish a feature data set of the trigger object; a position coordinate acquisition unit, the position coordinate acquisition unit is used to obtain the position coordinates of the dynamic tracking frame after the fall of the trigger object is authenticated by the feature data set; a relative distance calculation unit, the relative distance calculation unit is used to perform relative distance calculation using the position coordinates and the position of the visual sensor, and establish the automatic tracking focal length of the visual sensor based on the relative distance calculation result; a first visual acquisition result establishment unit, the first visual acquisition result establishment unit is used to establish the visual sensor acquisition frequency based on the feature data set and the ambient light state, perform tracking acquisition using the automatic tracking focal length and the visual sensor acquisition frequency, establish a first visual acquisition result, and use the first visual acquisition result as an acquisition dimension data of the first acquisition result.

[0053] Among them, a visual sensor acquisition frequency is established based on a feature data set and an ambient light state, tracking and acquisition are performed using automatic tracking focal length and a visual sensor acquisition frequency, a first visual acquisition result is established, and the first visual acquisition result is used as an acquisition dimension data of the first acquisition result. The first visual acquisition result establishment unit further includes: a first decision data acquisition subunit, the first decision data acquisition subunit is used to acquire the trigger object speed data in the feature data set, and use the trigger object speed data as the first decision data; a second decision data establishment subunit, the second decision data establishment subunit is used to acquire the trigger object brightness and the ambient light state in the feature data set to perform an environmental fusion evaluation of the trigger object, and establish the second decision data based on the environmental fusion evaluation result; a visual sensor acquisition frequency establishment subunit, the visual sensor acquisition frequency establishment subunit is used to establish the visual sensor acquisition frequency based on the first decision data and the second decision data.

[0054] Among them, the first visual acquisition result establishment unit further includes: an infrared sensor acquisition frequency establishment subunit, the infrared sensor acquisition frequency establishment subunit is used to establish the infrared sensor acquisition frequency using the trigger object speed data and the ambient light status; a second infrared acquisition result establishment subunit, the second infrared acquisition result establishment subunit is used to perform tracking and acquisition based on the infrared sensor acquisition frequency, and establish a second infrared acquisition result; a radar sensor acquisition frequency establishment subunit, the radar sensor acquisition frequency establishment subunit is used to call the environmental monitoring results, obtain radar signal interference, and establish the radar sensor acquisition frequency based on the radar signal interference and the trigger object speed data; a third radar acquisition result establishment subunit, the third radar acquisition result establishment subunit is used to perform tracking and acquisition based on the radar sensor acquisition frequency, and establish a third radar acquisition result; an acquisition dimension data acquisition subunit, the acquisition dimension data acquisition subunit is used to use the second infrared acquisition result and the third radar acquisition result as two acquisition dimension data of the first acquisition result.

[0055] The specific configuration of the trajectory backtracking fitting module 50 will be described in detail below. As described above, trajectory drawing is performed based on the first acquisition result and the second acquisition result, and trajectory backtracking fitting is performed using the trajectory backtracking network. The trajectory image of the trigger object is established according to the backtracking fitting result and the trajectory drawing result, and the high-altitude falling object alarm management is performed based on the trajectory image. The trajectory backtracking fitting module 50 further includes: an interpolation processing result establishment unit, the interpolation processing result establishment unit is used to call the interpolation channel of the trajectory backtracking network, and use the interpolation channel to interpolate the first acquisition result and the second acquisition result to establish the interpolation processing result; a backtracking fitting result establishment unit, the backtracking fitting result establishment unit is used to synchronize the interpolation processing result to the analysis layer of the trajectory backtracking network, perform backtracking fitting, and establish the backtracking fitting result; a trajectory image generation unit, the trajectory image generation unit is used to generate a trajectory image using the backtracking fitting result, wherein the trajectory image is marked with a throwing point.

[0056] Among them, the trajectory backtracking fitting module 50 further includes: a digital communication establishing unit, which is used to establish digital communication with the built-in camera of the target building; a prediction analysis result establishing unit, which is used to read the real-time captured data of the built-in camera, perform throwing tendency prediction analysis on the real-time captured data, and establish a prediction analysis result; a capture record generating unit, which is used to give a voice reminder based on the prediction analysis result and generate a capture record.

[0057] The high-altitude falling object monitoring device for smart city management provided by an embodiment of the present invention can execute the high-altitude falling object monitoring method for smart city management provided by any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.

[0058] Although the present application makes various references to certain modules in the device according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0059] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for monitoring falling objects from a high altitude for smart city management, characterized in that: The method comprises: Read the response accuracy of the high-altitude falling object monitoring, divide the target building into intervals based on the response accuracy, and configure the joint sensor using the interval division result and the response accuracy. The joint sensor includes a visual sensor, an infrared sensor, and a radar sensor; Using the environmental sensor to perform environmental monitoring at the location of the joint sensor, synchronizing the environmental monitoring results to the joint sensor, and performing acquisition initialization of the joint sensor; When the joint sensor is triggered, a dynamic tracking frame of the trigger object is established, and the automatic tracking focal length and acquisition frequency are configured with the dynamic tracking frame, tracking acquisition is performed, and a first acquisition result is established; Based on the interval division result, the joint sensor of the lower interval is placed in a scanning state, and after configuring the scanning focus point using the first acquisition result, the joint monitoring of the lower interval is performed to establish a second acquisition result; Based on the first acquisition result and the second acquisition result, trajectory drawing is performed, and trajectory backtracking fitting is performed using a trajectory backtracking network, a trajectory image of the trigger object is established according to the backtracking fitting result and the trajectory drawing result, and high-altitude falling object alarm management is performed based on the trajectory image; When the combined sensor is triggered, a dynamic tracking frame of the trigger object is established, which further includes: Using the acquisition initialization results, the response priority levels of the visual sensor, infrared sensor, and radar sensor in the joint sensor are configured, and the trigger thresholds are initialized based on the response priority levels. The trigger thresholds include the pixel change rate threshold, heat source displacement threshold, and speed change threshold of the visual sensor; The trigger threshold is used to perform fusion trigger identification on the real-time monitoring results of the joint sensor, the sensor with the highest response priority is called based on the fusion trigger identification result, and a dynamic tracking frame is established based on the sensor with the highest response priority and the fusion trigger identification result.

2. The method for monitoring falling objects from high altitude for smart city management according to claim 1, characterized in that: The step of configuring the dynamic tracking frame to automatically track the focal length and the acquisition frequency, performing tracking acquisition, and establishing a first acquisition result further includes: Collecting features of the trigger object based on the dynamic tracking frame to establish a feature data set of the trigger object; After the fall of the trigger object is authenticated using the feature data set, the position coordinates of the dynamic tracking frame are obtained; The relative distance is calculated using the position coordinates and the position of the visual sensor, and the automatic tracking focal length of the visual sensor is established based on the relative distance calculation result; A visual sensor acquisition frequency is established based on a feature data set and ambient light conditions, tracking acquisition is performed using automatic tracking focal length and visual sensor acquisition frequency, a first visual acquisition result is established, and the first visual acquisition result is used as an acquisition dimension data of the first acquisition result.

3. The method for monitoring falling objects from high altitude for smart city management as claimed in claim 2, characterized in that: The establishing of the visual sensor acquisition frequency based on the feature data set and the ambient light state further includes: Acquire the trigger object speed data in the feature data set, and use the trigger object speed data as the first decision data; Acquire the brightness of the trigger object and the ambient light state in the feature data set to perform an environmental fusion evaluation of the trigger object, and establish second decision data based on the environmental fusion evaluation result; A visual sensor acquisition frequency is established based on the first decision data and the second decision data.

4. The method for monitoring falling objects from high altitude for smart city management as claimed in claim 3, characterized in that: The method further comprises: Use the trigger object speed data and ambient light status to establish the infrared sensor acquisition frequency; Perform tracking acquisition based on the acquisition frequency of the infrared sensor to establish a second infrared acquisition result; Call the environmental monitoring results, obtain radar signal interference, and establish the radar sensor acquisition frequency based on radar signal interference and trigger object speed data; Perform tracking acquisition based on the radar sensor acquisition frequency and establish a third radar acquisition result; The second infrared acquisition result and the third radar acquisition result are used as two acquisition dimension data of the first acquisition result.

5. The method for monitoring falling objects from high altitude for smart city management according to claim 1, characterized in that: The method of performing trajectory backtracking fitting using the trajectory backtracking network and establishing a trajectory image of the trigger object according to the backtracking fitting result and the trajectory drawing result also includes: Calling the interpolation channel of the trajectory backtracking network, using the interpolation channel to perform interpolation processing on the first acquisition result and the second acquisition result, and establishing an interpolation processing result; The interpolation processing results are synchronized to the analysis layer of the trajectory backtracking network, backtracking fitting is performed, and backtracking fitting results are established; A trajectory image is generated using the backtracking fitting result, wherein the trajectory image is marked with ejection points.

6. The method for monitoring falling objects from high altitude for smart city management according to claim 1, characterized in that: The method further comprises: Establishing digital communications with the target building's built-in cameras; Reading the real-time captured data of the built-in camera, performing a prediction analysis on the real-time captured data, and establishing a prediction analysis result; Based on the prediction analysis results, voice reminders are given and capture records are generated.

7. A high-altitude falling object monitoring device for smart city management, characterized in that: The device is used to implement the high-altitude falling object monitoring method for smart city management according to any one of claims 1 to 6, and the device includes: A joint sensor configuration module, the joint sensor configuration module is used to read the response accuracy of high-altitude falling object monitoring, divide the target building into intervals based on the response accuracy, and configure the joint sensor using the interval division result and the response accuracy. The joint sensor includes a visual sensor, an infrared sensor, and a radar sensor; A collection initialization execution module, the collection initialization execution module is used to perform environmental monitoring of the joint sensor position using the environmental sensor, synchronize the environmental monitoring results to the joint sensor, and perform collection initialization of the joint sensor; A dynamic tracking frame establishment module, wherein the dynamic tracking frame establishment module is used to establish a dynamic tracking frame of the trigger object when the joint sensor is triggered, and configure the automatic tracking focal length and acquisition frequency with the dynamic tracking frame, perform tracking acquisition, and establish a first acquisition result; A joint monitoring execution module, the joint monitoring execution module is used to put the joint sensor of the lower interval into a scanning state based on the interval division result, configure the scanning focus point using the first acquisition result, perform joint monitoring of the lower interval, and establish a second acquisition result; A trajectory backtracking fitting module, wherein the trajectory backtracking fitting module is used to draw a trajectory based on the first acquisition result and the second acquisition result, and perform trajectory backtracking fitting using a trajectory backtracking network, establish a trajectory image of the trigger object according to the backtracking fitting result and the trajectory drawing result, and perform high-altitude falling object alarm management based on the trajectory image; The dynamic tracking frame establishment module also includes: a threshold trigger unit, which is used to configure the response priority level of the visual sensor, infrared sensor, and radar sensor in the joint sensor using the acquisition initialization result, and initialize the trigger threshold based on the response priority level, and the trigger threshold includes the pixel change rate threshold, heat source displacement threshold, and speed change threshold of the visual sensor; a fusion trigger recognition unit, which is used to use the trigger threshold to perform fusion trigger recognition on the real-time monitoring results of the joint sensor, call the sensor with the highest response priority level based on the fusion trigger recognition result, and establish a dynamic tracking frame based on the sensor with the highest response level and the fusion trigger recognition result.

Images