Intelligent target tracking method based on multifunctional loop fusion

Abstract

The invention discloses a target intelligent tracking method, device and equipment based on multifunctional loop fusion, a storage medium and an unmanned aerial vehicle, and the method comprises the steps: determining image motion feature data of a motion process according to a tracking video collected by a video tracking device of the unmanned aerial vehicle for the motion process of a target object, determining signal motion characteristic data of the motion process according to a tracking signal acquired by a signal tracking device of the unmanned aerial vehicle for the motion process; determining adjustment motion feature data of the motion process through reciprocal enhancement of the signal motion feature data to the image motion feature data; and controlling the flight state of the unmanned aerial vehicle according to iterative reasoning of the adjustment motion feature data and the signal motion feature data in the annular attractor network. Through information fusion of a reciprocal enhancement technology and an annular attractor network, accurate target tracking in a high-speed motion scene is realized, and the problems of high delay and large error when an unmanned aerial vehicle tracks a target object are solved.

Description

Technical Field

[0001] This application belongs to the field of unmanned aerial vehicles, and particularly relates to an intelligent target tracking method, device, equipment, storage medium and unmanned aerial vehicle based on multi-functional loop fusion. Background Art

[0002] In a complex forest environment, realizing the intelligent tracking of a target by an unmanned aerial vehicle has important application value. Especially in fields such as search and rescue, by combining remote target signals (such as Very High Frequency (VHF) signals) with visual obstacle avoidance technology, accurate tracking and obstacle avoidance of the target can be achieved. However, due to the complexity of the forest environment, target signals are vulnerable to interference from reflection, attenuation, and multipath effects, and the high-resolution computational complexity of visual obstacle avoidance increases the difficulty of realizing intelligent tracking.

[0003] Therefore, at present, a multi-method integration method that combines signal strength tracking and visual features is used to achieve target tracking.

[0004] However, although the existing multi-method integration improves the system adaptability, at high speeds, the signal strength tracking method has delays, and at the same time, the visual feature method is vulnerable to dynamic blur or occlusion, which leads to inaccurate prediction of the position of the target object and affects the target tracking effect. Summary of the Invention

[0005] This application aims to provide an intelligent target tracking method, device, equipment, storage medium and unmanned aerial vehicle based on multi-functional loop fusion, and at least solve the problems of high delay and large error when the unmanned aerial vehicle tracks a target object.

[0006] In a first aspect, an embodiment of this application discloses an intelligent target tracking method based on multi-functional loop fusion, which is applied to the controller of an unmanned aerial vehicle and includes: Determine the image motion feature data of the motion process according to the tracking video collected by the video tracking device of the unmanned aerial vehicle for the motion process of the target object, and determine the signal motion feature data of the motion process according to the tracking signal collected by the signal tracking device of the unmanned aerial vehicle for the motion process; Determine the adjusted motion feature data of the motion process through the reciprocal enhancement of the signal motion feature data to the image motion feature data; Control the flight state of the unmanned aerial vehicle according to the iterative reasoning of the adjusted motion feature data and the signal motion feature data in the circular attractor network.

[0007] In a second aspect, an embodiment of this application also discloses an unmanned aerial vehicle, including: Controller, video tracking device, and signal tracking device; The video tracking device is configured to collect a tracking video of the movement process of a target object; The signal tracking device is configured to collect a tracking signal of the movement process; The controller is configured to determine image motion feature data of the movement process according to the tracking video, and determine signal motion feature data of the movement process according to the tracking signal, and determine adjusted motion feature data of the movement process through reciprocal enhancement of the signal motion feature data to the image motion feature data, and control the flight state of the unmanned aerial vehicle according to iterative reasoning of the adjusted motion feature data and the signal motion feature data in a ring attractor network.

[0008] In a third aspect, an embodiment of the present application further discloses a target intelligent tracking device based on multi-functional loop fusion, which is applied to a controller of an unmanned aerial vehicle, and includes: A feature generation module, configured to determine image motion feature data of the movement process according to a tracking video collected by a video tracking device of the unmanned aerial vehicle for the movement process of a target object, and determine signal motion feature data of the movement process according to a tracking signal collected by a signal tracking device of the unmanned aerial vehicle for the movement process; A feature adjustment module, configured to determine adjusted motion feature data of the movement process through reciprocal enhancement of the signal motion feature data to the image motion feature data; A flight control module, configured to control the flight state of the unmanned aerial vehicle according to iterative reasoning of the adjusted motion feature data and the signal motion feature data in a ring attractor network.

[0009] In a fourth aspect, an embodiment of the present application further discloses an electronic device, including a processor and a memory, where the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, the steps of the method described in the first aspect are implemented.

[0010] In a fifth aspect, an embodiment of the present application further discloses a readable storage medium, where a program or instruction is stored on the readable storage medium, and when the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented.

[0011] In summary, in the embodiments of the present application, in complex environments and high-speed motion scenarios, the target area is highlighted in visual data by using the mutual benefit enhancement mechanism, and the signal and image data are combined, which improves the robustness and accuracy of the system and enhances the tracking effect. Furthermore, the process of a dynamic system is simulated through the annular attractor network to achieve multi-round iterative calculations to generate fused features, efficiently integrate multi-modal features, provide more accurate target tracking data, ensure that the unmanned aerial vehicle can adjust its flight state in real time, maintain stable tracking of the target, and ensure the real-time performance and adaptability of the unmanned aerial vehicle in complex dynamic scenarios. Thus, based on the method of the embodiments of the present application, under the conditions of resource constraints and complex environments, through the information fusion of the mutual benefit enhancement technology and the annular attractor network, accurate target tracking in high-speed motion scenarios is achieved, and the problems of high latency and large error when the unmanned aerial vehicle tracks the target object are solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In the drawings: Figure 1 FIG. is a flowchart of steps of a method for intelligent target tracking based on multi-functional loop fusion provided by an embodiment of the present application; Figure 2 FIG. is a flowchart of steps of another method for intelligent target tracking based on multi-functional loop fusion provided by an embodiment of the present application; Figure 3 FIG. is a data flow process under the embodiments of the present application; Figure 4 FIG. is a block diagram of a device for intelligent target tracking based on multi-functional loop fusion provided by an embodiment of the present application; Figure 5 FIG. is a schematic diagram of an unmanned aerial vehicle provided by an embodiment of the present application; Figure 6 FIG. is a block diagram of an electronic device of an embodiment provided by an embodiment of the present application; Figure 7 FIG. is a block diagram of an electronic device of another embodiment provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0014] The terms "first", "second", etc. in the description and claims of this application are used to distinguish similar objects, rather than to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of this application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of the same type, and do not limit the number of objects. For example, the first object can be one or more. In addition, "and / or" in the description and claims means at least one of the connected objects, and the character " / ", generally represents an "or" relationship between the associated objects before and after.

[0015] As Figure 1 shown, it is a target intelligent tracking method based on multi-functional loop fusion provided by an embodiment of this application.

[0016] The method may include the following steps: Step 101, determine the image motion feature data of the motion process according to the tracking video collected by the video tracking device of the unmanned aerial vehicle for the motion process of the target object, and determine the signal motion feature data of the motion process according to the tracking signal collected by the signal tracking device of the unmanned aerial vehicle for the motion process.

[0017] In some embodiments of this application, by determining the image motion feature data of the motion process according to the tracking video collected by the video tracking device of the unmanned aerial vehicle for the motion process of the target object, and determining the signal motion feature data of the motion process according to the tracking signal collected by the signal tracking device of the unmanned aerial vehicle for the motion process, visual information and signal information can be effectively combined to improve the accuracy of target tracking. During the execution of this step, the video tracking device collects the motion video of the target object, and at the same time, the signal tracking device collects the tracking signal of the target object, obtaining the image motion feature data and signal motion feature data during the motion process of the target object. The image motion feature data refers to the visual features of the target object in the motion process extracted from the video, and the signal motion feature data refers to the motion features of the target object extracted from the tracking signal. After executing this step, basic data combining visual and signal features is provided, providing a prerequisite for subsequent multi-modal feature fusion.

[0018] In a specific example, a drone applied to a search and rescue mission collects the motion video of a target object (such as a missing person) in the woods through its video tracking device, and determines the image motion feature data. At the same time, the signal tracking device of the drone collects the VHF signal of the target object and determines the signal motion feature data. After executing this process, the experimental personnel obtain the visual features and signal feature data of the target object during the motion process, which provide a key basis for the subsequent information fusion of mutual enhancement and the circular attractor network.

[0019] Step 102: Determine the adjusted motion feature data of the motion process through the reciprocal enhancement of the signal motion feature data and the image motion feature data.

[0020] In some embodiments of the present application, determining the adjusted motion feature data of the motion process through the reciprocal enhancement of the signal motion feature data and the image motion feature data can effectively improve the accuracy and robustness of target tracking. During the execution of this step, the system performs reciprocal enhancement on the signal motion feature data and the image motion feature data to generate the adjusted motion feature data. The signal motion feature data includes features such as the intensity and direction of the VHF signal, and the image motion feature data includes the visual features of the target object. After executing this step, by combining multi-modal features, the recognition and positioning ability of the target object is enhanced, providing more accurate data support for subsequent flight control.

[0021] In a specific example, in a drone search and rescue mission, the system performs reciprocal enhancement on the obtained VHF signal feature data and the image motion feature data of the target object. By combining the intensity and direction information of the VHF signal with the visual data, the adjusted motion feature data is generated. These adjusted feature data improve the recognition and positioning ability of the missing person, ensuring that the drone can track the target object more accurately in a complex forest environment and enhancing the success rate of the mission.

[0022] Step 103: Control the flight state of the unmanned aerial vehicle according to the iterative inference of the adjusted motion feature data and the signal motion feature data in the circular attractor network.

[0023] In some embodiments of the present application, by controlling the flight state of the unmanned aerial vehicle according to the iterative inference of the adjusted motion feature data and the signal motion feature data in the circular attractor network, precise flight control of the unmanned aerial vehicle can be achieved on the basis of multi-modal feature fusion. During the execution of this step, the system inputs the adjusted motion feature data and the signal motion feature data into the circular attractor network, and through iterative inference, generates fused features for controlling the flight state of the unmanned aerial vehicle. The circular attractor network is a neural network that realizes multi-round iterative calculations by simulating a dynamic system, and finally generates highly integrated feature data. After executing this step, the system can adjust the flight state of the unmanned aerial vehicle in real time to ensure stable tracking in a complex dynamic scenario.

[0024] In a specific example, in the UAV search and rescue mission, the system inputs the obtained adjusted motion feature data and signal motion feature data into the circular attractor network for iterative reasoning. Through multiple rounds of iterative calculations, fused feature data is generated to control the flight state of the UAV. After executing this process, the more accurate target tracking data obtained can be used to adjust the flight direction and speed of the UAV in real time, ensuring the stable tracking of the UAV in a complex forest environment and improving the success rate of the search and rescue mission.

[0025] In summary, in the embodiment of the present application, in a complex environment and high-speed motion scenario, by using the mutual reinforcement mechanism to highlight the target area in visual data, combining signal and image data, the robustness and accuracy of the system are improved, and the tracking effect is enhanced; furthermore, by simulating the process of a dynamic system through the circular attractor network, multiple rounds of iterative calculations are performed to generate fused features, efficiently integrating multi-modal features, providing more accurate target tracking data, ensuring that the unmanned aerial vehicle can adjust its flight state in real time, maintain stable tracking of the target, and ensure the real-time performance and adaptability of the unmanned aerial vehicle in a complex dynamic scenario. Therefore, based on the method of the embodiment of the present application, under the conditions of resource limitation and complex environment, through the information fusion of the mutual reinforcement technology and the circular attractor network, accurate target tracking in a high-speed motion scenario is achieved, and the problems of high latency and large error when the unmanned aerial vehicle tracks a target object are solved.

[0026] Figure 2 It is another target intelligent tracking method based on multi-functional loop fusion provided by the embodiment of the present application.

[0027] The method may include the following steps: Step 201, determine the image motion feature data of the motion process according to the tracking video collected by the video tracking device of the unmanned aerial vehicle for the motion process of the target object, and determine the signal motion feature data of the motion process according to the tracking signal collected by the signal tracking device of the unmanned aerial vehicle for the motion process.

[0028] The method shown in this step has been described in step 101 and will not be elaborated here.

[0029] Optionally, in order to determine the image motion feature data of the motion process according to the tracking video collected by the video tracking device of the unmanned aerial vehicle for the motion process of the target object, step 201 includes the following sub-steps: Sub-step 2011, extract multiple motion images of the target object changing with time from the tracking video, and determine multiple motion parameter data of the target object changing with time from the multiple motion images.

[0030] In some embodiments of the present application, by extracting multiple motion images of a target object changing over time from a tracking video and determining multiple motion parameter data of the target object changing over time from the multiple motion images, the dynamic changes of the target object can be effectively captured, and the accuracy of motion feature data can be improved. For example, a (4×4) convolutional kernel can be used to divide a motion image block into feature units, reducing the resolution to 1 / 4 of the original while increasing the number of feature channels to retain key feature information. During the execution of this step, the system extracts multiple motion images of the target object at different time points from the tracking video collected by the video tracking device of the unmanned aerial vehicle. These images will be used to analyze multiple parameter data of the target object during motion, such as position, speed, and direction. Motion parameter data refers to the numerical information reflecting the motion state of the target object extracted from multiple motion images. After executing this step, the system will obtain detailed motion parameter data of the target object changing over time, providing a basis for the determination of subsequent image motion feature data.

[0031] In a specific example, applied to the search and rescue mission of an unmanned aerial vehicle, the system extracts multiple images of a missing person moving in the woods from the tracking video collected by the video tracking device of the unmanned aerial vehicle. These images cover the motion states of the missing person at different moments. The experimenters determined multiple motion parameter data of the missing person changing over time, such as their position, moving speed, and direction, by analyzing these images. After executing this process, the obtained detailed motion parameter data will provide a key basis for further determining the image motion feature data, helping to improve the tracking accuracy and robustness of the unmanned aerial vehicle.

[0032] Sub-step 2012: Determine the image motion feature data according to the dynamic transformation of the multiple motion parameter data.

[0033] In some embodiments of the present application, by determining the image motion feature data according to the dynamic transformation of the multiple motion parameter data, the multiple motion parameter data can be further integrated for subsequent processing. During the execution of this step, the system first performs a dynamic transformation on the multiple motion parameter data to extract key data that can reflect the motion characteristics of the target object. The dynamic transformation includes processes such as dimensionality reduction, filtering, and feature enhancement of the motion parameter data. These processes will highlight and optimize the key information of the target object during motion. After executing this step, the obtained image motion feature data will have higher accuracy, providing an important basis for subsequent multi-modal feature fusion.

[0034] In a specific example, when applied to the search and rescue mission of drones, the system dynamically transforms multiple motion parameter data of the missing person (such as position, speed, direction, etc.). Through dimensionality reduction processing, the multi-dimensional motion parameter data is simplified into key data that can reflect the motion characteristics. At the same time, noise is removed through filtering, and the motion characteristics of the target object are highlighted through feature enhancement. After executing this process, the experimental personnel obtained high-precision image motion feature data, which provided key support for the subsequent mutual enhancement and information fusion of the ring attractor network, ensuring the precise tracking of the drone in a complex dynamic environment.

[0035] Optionally, sub-step 2012 includes the following sub-steps: Sub-step 20121, input multiple motion parameter data into the first differential network to obtain the output data of the dynamic system model of the target object through the global feature extraction of the image motion feature data by the first differential network.

[0036] In some embodiments of the present application, by inputting multiple motion parameter data into the first differential network to obtain the output data of the dynamic system model of the target object through the global feature extraction of the image motion feature data by the first differential network, the global features in the motion parameter data can be extracted, thereby enhancing the expression ability of the image motion feature data. During the execution of this step, the system inputs multiple motion parameter data into the first differential network, and the first differential network performs global feature extraction on these data to generate the output data of the dynamic system model of the target object. The output data of the dynamic system model is obtained by the neural network performing dynamic modeling of the motion parameter data for continuous time t: , where is the visual feature, is the global feature extraction function, implemented by a neural network, is the learning parameter, and thus the feature output is obtained. After executing this step, the system will obtain the output data of the dynamic system model containing global motion features, providing a basis for subsequent target enhancement.

[0037] In a specific example, when applied to the search and rescue mission of drones, the system inputs multiple motion parameter data of the missing person (such as position, speed, direction, etc.) into the first differential network. Through the global feature extraction function of this network, the system generates the output data of the dynamic system model reflecting the motion characteristics of the missing person. After executing this process, the obtained data containing global motion features will provide a key basis for the subsequent target enhancement of the image motion feature data, ensuring that the drone can accurately track the target object in a complex forest environment.

[0038] Sub-step 20122: Input the output data of the dynamic system model into the second differential network to enhance the target of the output data of the dynamic system model through the second differential network and obtain image motion feature data.

[0039] In some embodiments of the present application, by inputting the output data of the dynamic system model into the second differential network to enhance the target of the output data of the dynamic system model through the second differential network and obtain image motion feature data, the key features of the target object can be further extracted and enhanced, thereby improving the accuracy and robustness of the image motion feature data. During the execution of this step, the system inputs the output data of the dynamic system model obtained from the first differential network into the second differential network. The second differential network is a neural network-based model that highlights the key features of the target object through dynamic analysis and enhancement of the output data of the dynamic system model: , where is the output of the first layer, is the enhancement module, and the sparsity regularization term is used to suppress background noise, and are learning parameters. After executing this step, the system will obtain image motion feature data with high precision and high robustness, providing important support for subsequent multi-modal feature fusion and flight control of the unmanned aerial vehicle.

[0040] In a specific example, applied to the search and rescue mission of an unmanned aerial vehicle, the system inputs the output data of the dynamic system model of the missing person obtained from the first differential network into the second differential network. The experimenter uses the second differential network to perform dynamic analysis and enhancement on these data and extracts the key motion feature data of the missing person. These feature data include the motion direction, speed change, etc. of the missing person. After executing this process, the obtained high-precision image motion feature data will provide reliable data support for further information fusion and unmanned aerial vehicle flight control, improving the success rate of the search and rescue mission.

[0041] Optionally, the first differential network and the second differential network are jointly set in the first differential processing model.

[0042] In some embodiments of the present application, the first differential network and the second differential network are jointly arranged in the first differential processing model. Combining the two differential networks in one processing model can avoid multiple conversions and transmissions during data processing, reducing potential latency and error accumulation. At the same time, this structural design helps to optimize parameter sharing and model tuning in a unified architecture, improving the stability and performance of the overall algorithm. This can reduce system complexity and enhance the consistency and efficiency of data processing. By combining these two differential networks in the same processing model, global feature extraction and target enhancement can be effectively performed, simplifying the data transmission process and improving the utilization rate of computing resources.

[0043] Optionally, in order to determine the signal motion feature data of the motion process based on the tracking signals collected by the signal tracking device of the unmanned aerial vehicle during the motion process, step 201 includes the following sub-steps: Sub-step 2013: Extract multiple signal features of the tracking signal changing over time, and generate a signal sequence of the tracking signal based on the multiple signal features changing over time.

[0044] In some embodiments of the present application, by extracting multiple signal features of the tracking signal changing over time and generating a signal sequence of the tracking signal based on the multiple signal features changing over time, the dynamic change characteristics of the target object can be effectively captured, providing high-precision motion feature data. During the execution of this step, the system extracts multiple signal features of the target object at different times from the tracking signals collected by the signal tracking device of the unmanned aerial vehicle. These signal features will be used to generate the signal sequence of the target object, reflecting the feature changes of the target object during the motion process: The signal features can include signal strength (Received Signal Strength Indication, RSSI), direction information (such as the Angle of Arrival (AOA) of the signal), and frequency domain features (such as the output of the fast Fourier transform (FFT)), and encode them into a time series form such as : , where is the feature dimension, represents the n-dimensional feature component. After executing this step, the system will obtain the signal sequence of the target object, providing a basis for subsequent target enhancement and extraction of signal motion feature data.

[0045] In a specific example, applied to the drone search and rescue mission, the system collects the VHF signal emitted by the missing person from the signal tracking device of the drone, and extracts multiple signal characteristics of the target object changing over time, such as signal strength, direction, and frequency. By analyzing these signal characteristics, the experimenter generates a signal sequence of the target object, which reflects the dynamic changes of the missing person during the movement. After executing this process, the obtained signal sequence of the target object will provide a key basis for subsequent signal enhancement and extraction of signal motion feature data, helping to improve the tracking accuracy and robustness of the drone.

[0046] Sub-step 2014: Input the signal sequence into the third differential network to enhance the target of the signal sequence through the third differential network and obtain signal motion feature data.

[0047] In some embodiments of the present application, by inputting the signal sequence into the third differential network to enhance the target of the signal sequence through the third differential network and obtain signal motion feature data, the accuracy and robustness of the signal data can be improved, thereby improving the accuracy of target tracking. During the execution of this step, the system inputs the signal sequence generated in the previous step into the third differential network. The third differential network is a neural network-based model that extracts key motion feature data of the target object through dynamic analysis and enhancement of the input signal: , where is the sparsified signal feature, is the time series feature extraction function, is the sparsification function, and are learning parameters. These motion feature data can reflect the motion state of the target object in a complex dynamic environment. After executing this step, the system will obtain signal motion feature data with high precision and high robustness, providing key support for subsequent multi-modal feature fusion.

[0048] In a specific example, applied to the drone search and rescue mission, the system inputs the signal sequence generated from the VHF signal emitted by the missing person into the third differential network. The experimenter uses the third differential network to perform dynamic analysis and enhancement on the signal sequence, and extracts key motion feature data of the target object. These feature data include the motion direction of the missing person, speed change, etc. After executing this process, the obtained high-precision signal motion feature data will provide reliable data support for further information fusion and drone flight control, improving the success rate of the search and rescue mission.

[0049] Optionally, the third differential network is set in the second differential processing model.

[0050] In some embodiments of the present application, the third differential network is provided in the second differential processing model, which can enhance the modular design of the system and improve the efficiency and flexibility of data processing. By integrating the third differential network into the second differential processing model, independent processing and optimization can be effectively carried out, simplifying the system architecture and increasing the speed of data processing.

[0051] Step 202: Determine the adjusted motion feature data of the motion process by the reciprocal enhancement of the signal motion feature data to the image motion feature data.

[0052] The method shown in this step has been described in step 102 and will not be elaborated here.

[0053] Optionally, step 202 includes the following sub-steps: Sub-step 2021: Enhance the signal motion feature data to the image motion feature data according to a preset enhancement weight to obtain the feature adjustment input data of the image motion feature data.

[0054] In some embodiments of the present application, by enhancing the signal motion feature data to the image motion feature data according to a preset enhancement weight to obtain the feature adjustment input data of the image motion feature data, the motion features of the target object in two dimensions can be effectively fused. During the execution of this step, the system performs weighted fusion of the signal motion feature data and the image motion feature data according to the preset enhancement weight to generate the feature adjustment input data: , where is the enhancement weight for adjusting the feedback intensity. The enhancement weight refers to the weight values respectively assigned to the signal motion feature data and the image motion feature data during the data fusion process to ensure the balance and rationality after data fusion. After executing this step, the feature adjustment input data obtained by the system has higher applicability, providing key support for subsequent target enhancement.

[0055] In a specific example, applied to the UAV search and rescue mission, the system performs weighted fusion of the signal motion feature data (such as VHF signal features) and the image motion feature data (such as video tracking features) of the missing person according to the preset enhancement weight. Through this process, the experimental personnel generated the feature adjustment input data, which reflects the key features of the missing person during the motion process. After executing this process, the experimental personnel obtained the feature adjustment input data with high precision, which provided a reliable basis for subsequent multi-modal feature fusion and UAV flight control and improved the success rate of the search and rescue mission.

[0056] Sub-step 2022: Input the feature adjustment input data into the fourth differential network to obtain the adjusted motion feature data through the target enhancement of the fourth differential network to the feature adjustment input data.

[0057] In some embodiments of the present application, by inputting the feature-adjusted input data into the fourth differential network to enhance the target of the feature-adjusted input data through the fourth differential network, adjusted motion feature data can be obtained, which can further improve the accuracy and robustness of the feature data and ensure the accuracy and effectiveness of target tracking. During the execution of this step, the system inputs the feature-adjusted input data into the fourth differential network. The fourth differential network is a neural network-based model that extracts key motion feature data through multi-level dynamic analysis and enhancement of the input data. These motion feature data can reflect the motion state of the target object in a complex dynamic environment. After executing this step, the system will obtain adjusted motion feature data with high accuracy and high robustness, providing important support for subsequent multi-modal feature fusion and unmanned aerial vehicle flight control.

[0058] In a specific example, applied to the search and rescue mission of an unmanned aerial vehicle, the system inputs the feature-adjusted input data that combines signals and image features into the fourth differential network. Experimenters use these differential networks to perform dynamic analysis and enhancement on the feature-adjusted input data, and extract the key motion feature data of the missing person. These feature data include the motion direction and speed change of the missing person, etc. After executing this process, the obtained high-precision adjusted motion feature data will provide reliable data support for further information fusion and unmanned aerial vehicle flight control, and improve the success rate of the search and rescue mission.

[0059] It should be noted that in some embodiments of the present application, the second differential network can be used as the fourth differential network. By reusing the existing differential network, additional resource consumption and redundant network structures can be avoided, thereby improving the resource utilization rate of the system. At the same time, this shared usage method can ensure that feature enhancement in different processing stages is carried out in the same network, improving the consistency and accuracy of data processing and the overall performance of the system. In this way, resource sharing and optimization can be achieved, improving the system processing efficiency and consistency. By reusing the second differential network, the redundant design of the model can be reduced, and the overall architecture can be optimized, simplifying the implementation and maintenance of the system.

[0060] Step 203: By mapping the adjusted motion feature data and the signal motion feature data to different neurons of the ring attractor network respectively, input projection data for the ring attractor network is obtained.

[0061] In some embodiments of the present application, by mapping the adjusted motion feature data and the signal motion feature data to different neurons of the ring attractor network respectively to obtain input projection data for the ring attractor network, the efficiency and accuracy of feature integration can be improved in multi-modal feature fusion: , where , The projection matrix for visual and signal inputs. During the execution of this step, the system maps the adjusted motion feature data and the signal motion feature data to different neurons in the ring attractor network respectively, generating the input projection data for the ring attractor network. The ring attractor network is a neural network in which neurons are arranged in a closed loop, and the connection weights between neurons are defined by a Gaussian kernel function: , where i and j are neuron indices, is the diffusion parameter, controlling the connection diffusion range. After executing this step, the input projection data will become the initial state of information fusion in the ring attractor network, providing a basis for subsequent iterative reasoning.

[0062] In a specific example, in the UAV search and rescue mission, the system maps the adjusted motion feature data (such as visual feature data) and the signal motion feature data (such as VHF signal feature data) collected from the target object to different neurons in the ring attractor network respectively. These input projection data become the initial inputs in the ring attractor network, ensuring that the ring attractor network can efficiently integrate visual and signal features during multiple rounds of iterative reasoning, and finally generating more accurate target tracking data. Through these steps, the flight state of the UAV can be controlled more accurately, improving the success rate of the search and rescue mission.

[0063] Step 204, perform network dynamics iteration in the ring attractor network according to the input projection data to obtain the flight motion feature data of the unmanned aerial vehicle.

[0064] In some embodiments of the present application, by performing network dynamics iteration in the ring attractor network according to the input projection data to obtain the flight motion feature data of the unmanned aerial vehicle, multi-modal features can be efficiently integrated to generate accurate target tracking data. During the execution of this step, the system uses the input projection data and performs multiple rounds of iterative calculations through the ring attractor network to generate the flight motion feature data: , where is the fusion feature state of the ring attractor network at time t, and the initial time is 0. is the connection matrix of the ring network. After executing this step, the system can generate accurate flight motion feature data, providing strong data support for the flight control of the unmanned aerial vehicle.

[0065] In a specific example, applied to the search and rescue mission of an unmanned aerial vehicle (UAV), the system uses the visual features and signal features collected from the target object as input projection data, and performs multiple rounds of iterative calculations through a ring attractor network. This network dynamically adjusts the connection weights between neurons, fuses the input projection data, and generates the flight motion feature data of the UAV. After executing this process, the obtained accurate flight motion feature data is used to control the flight direction and speed of the UAV, ensuring the precise tracking of the UAV in a complex forest environment and improving the success rate of the search and rescue mission.

[0066] Step 205: Generate flight control data for the unmanned aerial vehicle according to the flight motion feature data, and control the flight state of the unmanned aerial vehicle according to the flight control data.

[0067] In some embodiments of the present application, by generating flight control data for the unmanned aerial vehicle according to the flight motion feature data and controlling the flight state of the unmanned aerial vehicle according to the flight control data, it can be ensured that the unmanned aerial vehicle finally obtains real-time control information. During the execution of this step, the system first generates flight control data for the unmanned aerial vehicle according to the obtained flight motion feature data. The flight motion feature data is accurate feature data obtained through the iterative reasoning of the ring attractor network, and the flight control data is data used to control the flight direction and speed of the unmanned aerial vehicle. For example, the fused features are projected into a two-dimensional target space through a dimensionality reduction module to generate the outputs of roll and yaw angles: , where Roll and Yaw are the roll angle and yaw angle outputs of the unmanned aerial vehicle respectively, used to control the flight direction adjustment of the unmanned aerial vehicle. After executing this step, the system uses the generated flight control data to real-time control the flight state of the unmanned aerial vehicle, ensuring precise tracking in a complex dynamic environment.

[0068] In a specific example, in the search and rescue mission of a UAV, the experimenter uses the flight motion feature data obtained through the iterative reasoning of the ring attractor network to generate the flight control data of the UAV. These flight control data include information such as the flight direction, speed, and altitude of the UAV. The experimenter then adjusts the flight state of the UAV in real-time according to these flight control data, enabling the UAV to stably track the target object (such as a missing person) in a complex forest environment. Through this process, the experimenter can ensure the precise tracking of the UAV in a dynamic environment and improve the success rate of the search and rescue mission.

[0069] As Figure 3 shown, it is a data flow process under the embodiments of the present application: S1. Image dimensionality reduction input: Extract the features of the image by reducing the resolution of the image through convolutional pooling while retaining important feature information; S2. Tracking signal encoding input: Convert the original signal into feature data that can be further processed; S3. Feature transformation mapping: Use a two-layer Ordinary Differential Equation (ODE) model to perform feature transformation mapping on video data. The first layer is responsible for global feature extraction, and the second layer is responsible for target enhancement; S4. Feature sparsification: Use a single-layer ODE model for feature sparsification to enhance the target of signal data processing; S5. Visual reciprocal enhancement: Achieve visual reciprocal enhancement through data fusion, use signal data to enhance visual features, and make the target area more obvious; S6. Feature transformation mapping: Use the second layer of the two-layer ODE model again for feature transformation mapping, aiming to further optimize the target feature data; S7. Visual and signal input connection: Connect the visual input and the signal input and input them into the recurrent attractor network; S8. Iterative calculation and solution: Perform iterative calculation and solution through the recurrent attractor network. Through multiple rounds of iteration, generate fused feature data; S9. Roll and yaw output: Perform projection through a Multilayer Perceptron (MLP) process, convert the fused feature data into control commands to adjust the flight state of the unmanned aerial vehicle.

[0070] In summary, in the embodiment of the present application, in complex environments and high-speed motion scenarios, the reciprocal enhancement mechanism is used to highlight the target area in visual data, combine signal and image data, improve the robustness and accuracy of the system, and enhance the tracking effect; furthermore, the process of a dynamic system is simulated through the recurrent attractor network, multi-round iterative calculations are performed to generate fused features, multi-modal features are efficiently integrated, and more accurate target tracking data is provided to ensure that the unmanned aerial vehicle can adjust its flight state in real time, maintain stable tracking of the target, and ensure the real-time performance and adaptability of the unmanned aerial vehicle in complex dynamic scenarios. Therefore, based on the method of the embodiment of the present application, under the conditions of resource constraints and complex environments, through the information fusion of the reciprocal enhancement technology and the recurrent attractor network, accurate target tracking in high-speed motion scenarios is achieved, and the problems of high latency and large error when the unmanned aerial vehicle tracks the target object are solved.

[0071] Reference Figure 4 , which shows a target intelligent tracking device 30 provided by the embodiment of the present application, applied to the controller of an unmanned aerial vehicle, including: A feature generation module 301, configured to determine image motion feature data of the motion process according to the tracking video collected by the video tracking device of the unmanned aerial vehicle for the motion process of the target object, and determine signal motion feature data of the motion process according to the tracking signal collected by the signal tracking device of the unmanned aerial vehicle for the motion process; A feature adjustment module 302, configured to determine adjusted motion feature data of the motion process through reciprocal enhancement of the signal motion feature data to the image motion feature data; A flight control module 303, configured to control the flight state of the unmanned aerial vehicle according to iterative reasoning of the adjusted motion feature data and the signal motion feature data in the annular attractor network.

[0072] Optionally, the feature generation module 301 includes: A motion extraction sub-module, configured to extract multiple motion images of the target object changing with time from the tracking video, and determine multiple motion parameter data of the target object changing with time from the multiple motion images; A dynamic transformation sub-module, configured to determine the image motion feature data according to the dynamic transformation of the multiple motion parameter data.

[0073] Optionally, the dynamic transformation sub-module includes: A first differential unit, configured to input the multiple motion parameter data into a first differential network, so as to obtain output data of the dynamic system model of the target object through global feature extraction of the image motion feature data by the first differential network; A second differential unit, configured to input the output data of the dynamic system model into a second differential network, so as to obtain the image motion feature data through target enhancement of the output data of the dynamic system model by the second differential network.

[0074] Optionally, the feature generation module 301 includes: A signal sequence sub-module, configured to extract multiple signal features of the tracking signal changing with time, and generate a signal sequence of the tracking signal according to the multiple signal features changing with time; A signal feature sub-module, configured to input the signal sequence into a third differential network, so as to obtain the signal motion feature data through target enhancement of the signal sequence by the third differential network.

[0075] Optionally, the feature adjustment module 302 includes: A weighting sub-module, configured to enhance the signal motion feature data to the image motion feature data according to a preset enhancement weight to obtain feature adjustment input data of the image motion feature data; An adjustment sub-module, configured to input the feature adjustment input data into a fourth differential network, so as to obtain the adjusted motion feature data through target enhancement of the feature adjustment input data by the fourth differential network.

[0076] Optionally, the flight control module 303 includes: An input sub-module, configured to obtain input projection data for the ring attractor network by respectively mapping the adjusted motion feature data and the signal motion feature data to different neurons of the ring attractor network; An iteration sub-module, configured to perform network dynamics iteration in the ring attractor network according to the input projection data to obtain flight motion feature data of the unmanned aerial vehicle; A control sub-module, configured to generate flight control data for the unmanned aerial vehicle according to the flight motion feature data, and control the flight state of the unmanned aerial vehicle according to the flight control data.

[0077] Reference Figure 5 , which shows an unmanned aerial vehicle M provided by an embodiment of the present application, including: A controller (not shown in the figure), a video tracking device X, and a signal tracking device Y; The video tracking device X is configured to collect and track a video of the motion process of the target object O; The signal tracking device Y is configured to collect and track a signal during the motion process; The controller is configured to determine image motion feature data of the motion process according to the tracking video, determine signal motion feature data of the motion process according to the tracking signal, determine adjusted motion feature data of the motion process through the reciprocal enhancement of the signal motion feature data to the image motion feature data, and control the flight state of the unmanned aerial vehicle M according to the iterative reasoning of the adjusted motion feature data and the signal motion feature data in the ring attractor network.

[0078] It should be emphasized that Figure 5 the external shapes and structures of the unmanned aerial vehicle M, the video tracking device X, and the signal tracking device Y in this application are all exemplary, and the external shapes and structures of these devices can be adjusted and optimized according to specific application requirements. The external shapes and structures of the above devices are only for illustrative purposes and should not be regarded as a limitation on the protection scope of the present invention. The protection scope of the present application shall be subject to the claims and has nothing to do with the specific external shapes of the devices.

[0079] In summary, in the embodiments of the present application, in complex environments and high-speed motion scenarios, the target area is highlighted in visual data by using the mutual benefit enhancement mechanism, and the signal and image data are combined, which improves the robustness and accuracy of the system and enhances the tracking effect. Furthermore, the process of a dynamic system is simulated through a ring attractor network to implement multi-round iterative calculations to generate fused features, efficiently integrate multi-modal features, provide more accurate target tracking data, ensure that the unmanned aerial vehicle can adjust its flight state in real time, maintain stable tracking of the target, and ensure the real-time performance and adaptability of the unmanned aerial vehicle in complex dynamic scenarios. Thus, based on the method of the embodiments of the present application, under the conditions of resource constraints and complex environments, through the information fusion of the mutual benefit enhancement technology and the ring attractor network, accurate target tracking in high-speed motion scenarios is achieved, and the problems of high latency and large error when the unmanned aerial vehicle tracks a target object are solved.

[0080] Referring Figure 6 , the electronic device 500 may include one or more of the following components: a processing component 502, a memory 504, a power supply component 506, a multimedia component 508, an audio component 510, an input / output (I / O) interface 512, a sensor component 514, and a communication component 516.

[0081] The processing component 502 generally controls the overall operation of the electronic device 500, such as operations associated with display, telephone calls, data communication, camera operations, and recording operations. The processing component 502 may include one or more processors 520 to execute instructions to complete all or part of the steps of the above methods. In addition, the processing component 502 may include one or more modules to facilitate the interaction between the processing component 502 and other components. For example, the processing component 502 may include a multimedia module to facilitate the interaction between the multimedia component 508 and the processing component 502.

[0082] The memory 504 is used to store various types of data to support the operation of the electronic device 500. Examples of these data include instructions for any application or method operating on the electronic device 500, contact data, phone book data, messages, pictures, multimedia, etc. The memory 504 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disc.

[0083] The power supply component 506 provides power to various components of the electronic device 500. The power supply component 506 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the electronic device 500.

[0084] The multimedia component 508 includes an interface that provides an output interface between the electronic device 500 and the user. In some embodiments, the interface may include a liquid crystal display (LCD) and a touch panel (TP). If the interface includes a touch panel, the interface can be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also detect the duration and pressure associated with the touch or swipe operation. In some embodiments, the multimedia component 508 includes a front camera and / or a rear camera. When the electronic device 500 is in an operating mode, such as a shooting mode or a multimedia mode, the front camera and / or the rear camera can receive external multimedia data. Each of the front camera and the rear camera can be a fixed optical lens system or have a focal length and optical zoom capabilities.

[0085] The audio component 510 is used to output and / or input audio signals. For example, the audio component 510 includes a microphone (MIC) that is used to receive external audio signals when the electronic device 500 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals can be further stored in the memory 504 or transmitted via the communication component 516. In some embodiments, the audio component 510 further includes a speaker for outputting audio signals.

[0086] The input / output I / O interface 512 provides an interface between the processing component 502 and a peripheral interface module, and the peripheral interface module can be a keyboard, a click wheel, buttons, etc. These buttons can include but are not limited to: a home button, a volume button, a power button, and a lock button.

[0087] The sensor component 514 includes one or more sensors for providing a status assessment of various aspects of the electronic device 500. For example, the sensor component 514 can detect the on / off state of the electronic device 500, the relative positioning of components, such as the display and the keypad of the electronic device 500. The sensor component 514 can also detect a change in the position of the electronic device 500 or a component of the electronic device 500, the presence or absence of user contact with the electronic device 500, the orientation or acceleration / deceleration of the electronic device 500, and the temperature change of the electronic device 500. The sensor component 514 can include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor component 514 can also include a light sensor, such as a CMOS or a CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 514 can further include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0088] The communication component 516 is used to facilitate communication between the electronic device 500 and other devices in a wired or wireless manner. The electronic device 500 can access a communication standard-based wireless network, such as WiFi, a carrier network (such as 2G, 3G, 4G, or 5G), or a combination thereof. In an exemplary embodiment, the communication component 516 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 516 further includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0089] In an exemplary embodiment, the electronic device 500 can be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for implementing the methods provided in the embodiments of the present application.

[0090] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 504 including instructions, and the above instructions can be executed by a processor 520 of the electronic device 500 to complete the above methods. For example, the non-transitory storage medium can be a ROM, Random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0091] Figure 7 is a block diagram of an electronic device 600 according to another embodiment of the present invention. For example, the electronic device 600 can be provided as a server.

[0092] Refer to Figure 7 , the electronic device 600 includes a processing component 622, which further includes one or more processors, and memory resources represented by a memory 632 for storing instructions executable by the processing component 622, such as application programs. The application programs stored in the memory 632 can include one or more modules each corresponding to a set of instructions. In addition, the processing component 622 is configured to execute instructions to perform the methods provided in the embodiments of the present application.

[0093] The electronic device 600 may further include a power supply component 626 configured to perform power management of the electronic device 600, a wired or wireless network interface 650 configured to connect the electronic device 600 to a network, and an input / output (I / O) interface 658. The electronic device 600 may operate based on an operating system stored in the memory 632, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSD TM or the like.

[0094] It should be noted that, for the method embodiments of the present application, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the embodiments of the present application are not limited by the described action sequence, because according to the embodiments of the present application, some steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present application.

[0095] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the application disclosed herein. The present application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only to be regarded as exemplary, and the true scope and spirit of the present application are pointed out by the following claims.

[0096] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A target intelligent tracking method based on multi-function loop fusion, characterized in that: Controllers for unmanned aerial vehicles include: Determine image motion feature data of the motion process based on a tracking video collected by a video tracking device of the unmanned aerial vehicle during the motion process of the target object, and determine signal motion feature data of the motion process based on a tracking signal collected by a signal tracking device of the unmanned aerial vehicle during the motion process; Determine the adjusted motion feature data of the motion process by reciprocally enhancing the image motion feature data with the signal motion feature data; The flight state of the unmanned aerial vehicle is controlled according to iterative reasoning of the adjustment motion feature data and the signal motion feature data in the ring attractor network.

2. The target intelligent tracking method based on multifunctional loop fusion as claimed in claim 1, characterized in that: The step of determining the image motion feature data of the motion process based on the tracking video collected by the video tracking device of the unmanned aerial vehicle during the motion process of the target object comprises: Extracting a plurality of motion images of the target object changing over time from the tracking video, and determining a plurality of motion parameter data of the target object changing over time from the plurality of motion images; The image motion characteristic data is determined based on the dynamic transformation of the plurality of motion parameter data.

3. The target intelligent tracking method based on multi-function loop fusion as claimed in claim 2, characterized in that: Determining the image motion feature data according to the dynamic transformation of the plurality of motion parameter data comprises: Inputting the plurality of motion parameter data into a first differential network, so as to extract global features of the image motion feature data through the first differential network, and obtain output data of the dynamic system model of the target object; The dynamic system model output data is input into a second differential network, so as to target-enhance the dynamic system model output data through the second differential network to obtain the image motion feature data.

4. The target intelligent tracking method based on multi-function loop fusion as claimed in claim 1, characterized in that: The step of determining the signal motion characteristic data of the motion process based on the tracking signal collected by the signal tracking device of the unmanned aerial vehicle during the motion process comprises: Extracting a plurality of signal features of the tracking signal that vary with time, and generating a signal sequence of the tracking signal according to the plurality of signal features that vary with time; The signal sequence is input into a third differential network so as to enhance the target of the signal sequence through the third differential network and obtain the signal motion characteristic data.

5. The target intelligent tracking method based on multi-function loop fusion as claimed in claim 1, characterized in that: The reciprocal enhancement of the image motion feature data by the signal motion feature data to determine the adjusted motion feature data of the motion process includes: enhancing the signal motion feature data to the image motion feature data according to a preset enhancement weight to obtain feature adjustment input data of the image motion feature data; The feature-adjusted input data is input into a fourth differential network, so as to perform target enhancement on the feature-adjusted input data through the fourth differential network to obtain the adjusted motion feature data.

6. The target intelligent tracking method based on multi-function loop fusion as claimed in claim 1, characterized in that: The method of controlling the flight state of the unmanned aerial vehicle according to the iterative reasoning of the adjustment motion characteristic data and the signal motion characteristic data in the ring attractor network includes: By mapping the adjustment motion feature data and the signal motion feature data to different neurons of the annular attractor network respectively, so as to obtain input projection data to the annular attractor network; Performing network dynamics iteration in the annular attractor network according to the input projection data to obtain flight motion characteristic data of the unmanned aerial vehicle; The flight control data of the unmanned aerial vehicle is generated according to the flight motion characteristic data, and the flight state of the unmanned aerial vehicle is controlled according to the flight control data.

7. An unmanned aerial vehicle, characterized in that: include: controller, video tracking device and signal tracking device; The video tracking device is used to collect tracking video of the movement process of the target object; The signal tracking device is used to collect tracking signals for the motion process; The controller is used to determine the image motion feature data of the motion process based on the tracking video, and to determine the signal motion feature data of the motion process based on the tracking signal, and to determine the adjustment motion feature data of the motion process through reciprocal enhancement of the image motion feature data by the signal motion feature data, and to control the flight state of the unmanned aerial vehicle based on iterative reasoning of the adjustment motion feature data and the signal motion feature data in a ring attractor network.

8. A target intelligent tracking device based on multi-function loop fusion, characterized in that: Controllers for unmanned aerial vehicles include: A feature generation module, used to determine image motion feature data of the motion process based on a tracking video collected by a video tracking device of the unmanned aerial vehicle during the motion process of the target object, and to determine signal motion feature data of the motion process based on a tracking signal collected by a signal tracking device of the unmanned aerial vehicle during the motion process; A feature adjustment module, configured to determine the adjusted motion feature data of the motion process by reciprocally enhancing the image motion feature data through the signal motion feature data; A flight control module is used to control the flight state of the unmanned aerial vehicle according to iterative reasoning of the adjustment motion characteristic data and the signal motion characteristic data in a ring attractor network.

9. An electronic device, characterized in that: include: A processor, a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the steps of the target intelligent tracking method based on multi-function loop fusion as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the steps of the target intelligent tracking method based on multi-function loop fusion as described in any one of claims 1 to 6.

Images