Shared electric bicycle-oriented violent destruction behavior identification method, system and equipment

Abstract

The invention provides a violent damage behavior identification method, system and equipment for a shared electric bicycle. The method comprises the following steps: acquiring multi-modal monitoring data of the shared electric bicycle; carrying out preprocessing on the multi-mode monitoring data; performing feature extraction on the preprocessed multi-modal monitoring data to obtain a multi-modal monitoring feature vector; inputting the multi-modal monitoring feature vectors into a pre-constructed violent destruction recognition model, fusing the multi-modal monitoring feature vectors by the violent destruction recognition model to obtain a fused feature vector, judging whether a violent destruction behavior for the shared electric bicycle exists or not according to the fused feature vector, and if yes, judging whether the violent destruction behavior for the shared electric bicycle exists or not. If yes, determining the type of the violent destruction behavior corresponding to the multi-modal monitoring data; wherein the violent damage identification model is obtained by training a sample multi-modal monitoring feature vector corresponding to sample multi-modal monitoring data of the sample shared electric bicycle. According to the invention, the violent destruction behavior for the shared electric bicycle can be accurately identified.

Description

Technical Field

[0001] The present invention relates to the technical field of shared electric motorcycles, and in particular to a method, system and device for identifying violent destructive behavior on shared electric motorcycles. Background Art

[0002] Shared electric motorcycles are an important option for short-distance travel in cities, providing citizens with a convenient and environmentally friendly way to travel. However, with the popularity of shared electric motorcycles, phenomena such as violent vandalism and malicious damage have occurred frequently, not only affecting the normal travel of citizens, but also causing huge economic losses to shared travel companies. Therefore, developing a method and system that can effectively identify and prevent violent vandalism of shared electric motorcycles is of great significance for protecting public property safety and improving the quality of shared travel services in cities. The existing methods for identifying violent vandalism of shared electric motorcycles only consider a single factor, and the recognition accuracy is low. Summary of the Invention

[0003] The present invention provides a method, system and device for identifying violent destructive behavior on shared electric motorcycles, which are used to solve the defect that the existing method for identifying violent destructive behavior on shared electric motorcycles only considers a single factor and has low recognition accuracy.

[0004] The present invention provides a method for identifying violent destructive behavior on shared electric motorcycles, comprising: Obtain multimodal monitoring data of shared electric motorcycles; Preprocessing the multimodal monitoring data to obtain preprocessed multimodal monitoring data; Performing feature extraction on the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; Inputting the multimodal monitoring feature vector into a pre-built violent vandalism identification model, the violent vandalism identification model fuses the multimodal monitoring feature vector to obtain a fused feature vector. Based on the fused feature vector, it is determined whether there is violent vandalism directed at the shared electric motorcycle. If so, the category of violent vandalism corresponding to the multimodal monitoring data is determined; Among them, the violent destruction identification model is obtained by training based on the sample multimodal monitoring feature vector corresponding to the sample multimodal monitoring data of the sample shared electric motorcycle.

[0005] In some embodiments, the multimodal monitoring data includes image data, sound data and / or vibration data; the multimodal monitoring feature vector includes an image feature vector, a sound feature vector and / or a vibration feature vector.

[0006] In some embodiments, fusing the multimodal monitoring feature vectors to obtain a fused feature vector includes: determining a weight of the image feature vector, a weight of the sound feature vector, and / or a weight of the vibration feature vector; The image feature vector, the sound feature vector and / or the vibration feature vector are fused according to the weight of the image feature vector, the weight of the sound feature vector and / or the weight of the vibration feature vector to obtain the fused feature vector.

[0007] In some embodiments, the method further comprises: Generate early warning information and issue an alarm based on the category of the violent and destructive behavior; Determine the abnormal user who has committed the violent destructive behavior and mark the abnormal user.

[0008] In some embodiments, the training process of the brute force damage recognition model includes: Obtain sample multimodal monitoring data of sample shared electric motorcycles; Preprocessing the sample multimodal monitoring data to obtain preprocessed sample multimodal monitoring data; Performing feature extraction on the preprocessed sample multimodal monitoring data to obtain a sample multimodal monitoring feature vector; Determining a category label of the violent destructive behavior corresponding to the sample multimodal monitoring data; The sample multimodal monitoring feature vector is used as a training sample, and the category label of the violent destruction behavior corresponding to the sample multimodal monitoring data is used as a sample label to train an initial violent destruction recognition model. After the training is completed, the violent destruction recognition model is obtained.

[0009] In some embodiments, the sample multimodal monitoring data includes sample image data, sample sound data and / or sample vibration data; the sample multimodal monitoring feature vector includes a sample image feature vector, a sample sound feature vector and / or a sample vibration feature vector.

[0010] The present invention also provides a system for identifying violent and destructive behaviors of shared electric motorcycles, comprising: An acquisition unit, used to acquire multimodal monitoring data of shared electric motorcycles; a preprocessing unit, configured to preprocess the multimodal monitoring data to obtain preprocessed multimodal monitoring data; A feature extraction unit, configured to extract features from the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; an identification unit, configured to input the multimodal monitoring feature vector into a pre-built violent vandalism identification model, have the violent vandalism identification model fuse the multimodal monitoring feature vector to obtain a fused feature vector, and determine, based on the fused feature vector, whether there is violent vandalism directed at the shared electric motorcycle; if so, determine the category of violent vandalism corresponding to the multimodal monitoring data; Among them, the violent destruction identification model is obtained by training based on the sample multimodal monitoring feature vector corresponding to the sample multimodal monitoring data of the sample shared electric motorcycle.

[0011] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the program, it implements any of the above-described methods for identifying violent destructive behaviors for shared electric motorcycles.

[0012] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for identifying violent destructive behavior for shared electric motorcycles as described in any of the above methods is implemented.

[0013] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for identifying violent destructive behaviors for shared electric motorcycles.

[0014] The method, system and device for identifying violent destructive behavior towards shared electric motorcycles provided by the present invention obtain multimodal monitoring data of shared electric motorcycles and preprocess the multimodal monitoring data; perform feature extraction on the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; input the multimodal monitoring feature vector into a pre-built violent destructive behavior recognition model, and the violent destructive behavior recognition model fuses the multimodal monitoring feature vector to obtain a fused feature vector. Based on the fused feature vector, it is determined whether there is violent destructive behavior towards shared electric motorcycles. If so, the category of the violent destructive behavior corresponding to the multimodal monitoring data is determined, thereby improving the accuracy of violent destructive behavior recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1This is a flow chart of a method for identifying violent destructive behavior on shared electric motorcycles provided in an embodiment of the present invention.

[0017] Figure 2 4 is a flowchart of a training process of a violent destruction recognition model provided by an embodiment of the present invention.

[0018] Figure 3 2 is a schematic diagram of the structure of a violent destructive behavior identification system for shared electric motorcycles provided by an embodiment of the present invention.

[0019] Figure 4 It is a structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0021] Figure 1 The flowchart of the method for identifying violent destructive behavior on shared electric motorcycles provided by the embodiment of the present invention is as follows. Figure 1 As shown, a method for identifying violent destructive behavior of shared electric motorcycles is provided, comprising the following steps: step 110, step 120, step 130, and step 140. The steps of the method flow are merely a possible implementation of the present invention.

[0022] Step 110: Obtain multimodal monitoring data of shared electric motorcycles.

[0023] Optionally, multimodal monitoring data is collected through monitoring equipment on the shared electric motorcycle and transmitted to a cloud server; the monitoring equipment includes high-definition cameras, sound sensors, vibration sensors, etc.

[0024] It should be noted that when someone tries to damage a shared electric motorcycle, the high-definition camera can capture this behavior and provide evidence for subsequent accountability; when someone tries to damage a shared electric motorcycle, abnormal noise may be generated, and the sound sensor can capture this signal; when someone tries to damage a shared electric motorcycle by knocking, bumping, etc., a vibration signal will be generated, and the vibration sensor can capture this signal.

[0025] In some embodiments, the multimodal monitoring data includes image data, sound data, and / or vibration data.

[0026] Optionally, surveillance video is collected by a high-definition camera on a shared electric motorcycle, and the surveillance video is processed to obtain image data and sound data.

[0027] Step 120: preprocess the multimodal monitoring data to obtain preprocessed multimodal monitoring data.

[0028] Optionally, the multimodal monitoring data is preprocessed by data cleaning, data enhancement, normalization, denoising, etc.

[0029] Step 130: Perform feature extraction on the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector.

[0030] In some embodiments, the multimodal monitoring feature vector includes an image feature vector, a sound feature vector, and / or a vibration feature vector.

[0031] Among them, the image feature vector includes the facial features of the violent vandal, the violent damage area features of the shared electric motorcycle, etc.; the sound feature vector includes the sound features of the violent vandal, the noise features of the shared electric motorcycle, etc.; the vibration feature vector includes the vibration frequency, vibration amplitude and other features of the shared electric motorcycle.

[0032] Optionally, feature extraction is performed on the image data to obtain an image feature vector; feature extraction is performed on the sound data to obtain a sound feature vector; and feature extraction is performed on the vibration data to obtain a vibration feature vector.

[0033] Step 140: Input the multimodal monitoring feature vector into a pre-built violent destruction recognition model. The violent destruction recognition model fuses the multimodal monitoring feature vector to obtain a fused feature vector. Based on the fused feature vector, determine whether there is violent destruction behavior against shared electric motorcycles. If so, determine the category of violent destruction behavior corresponding to the multimodal monitoring data.

[0034] Among them, the violent destruction recognition model is trained based on the sample multimodal monitoring feature vector corresponding to the sample multimodal monitoring data of the sample shared electric motorcycle.

[0035] Optionally, the image feature vector, sound feature vector and / or vibration feature vector are input into a violent destruction recognition model, which fuses the image feature vector, sound feature vector and / or vibration feature vector to obtain a fused feature vector. Based on the fused feature vector, it is determined whether there is any violent destruction behavior against shared electric motorcycles. If so, the category of the violent destruction behavior corresponding to the multimodal monitoring data is determined.

[0036] Optionally, the violent destruction recognition model includes a feature fusion layer, a recognition layer, and a classification layer.

[0037] Optionally, the image feature vector, the sound feature vector and / or the vibration feature vector are input into the feature fusion layer to obtain a fused feature vector output by the feature fusion layer.

[0038] Optionally, the fused feature vector is input into the recognition layer to obtain the violent destructive behavior recognition result output by the recognition layer, that is, whether the violent destructive behavior exists or not.

[0039] Optionally, when it is determined that there is violent destructive behavior, the fused feature vector is input into the classification layer to obtain the category of violent destructive behavior towards shared electric motorcycles output by the classification layer.

[0040] In some embodiments, fusing the multimodal monitoring feature vectors to obtain a fused feature vector includes: determining a weight of an image feature vector, a weight of a sound feature vector, and / or a weight of a vibration feature vector; The image feature vector, the sound feature vector and / or the vibration feature vector are fused according to the weight of the image feature vector, the weight of the sound feature vector and / or the weight of the vibration feature vector to obtain a fused feature vector.

[0041] It can be understood that by determining the weight of the image feature vector, the weight of the sound feature vector and / or the weight of the vibration feature vector, and fusing the image feature vector, the sound feature vector and / or the vibration feature vector on this basis, the accuracy of feature fusion is improved, which helps to improve the accuracy of identifying violent and destructive behaviors.

[0042] In an embodiment of the present invention, multimodal monitoring data of shared electric motorcycles is obtained and preprocessed; feature extraction is performed on the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; the multimodal monitoring feature vector is input into a pre-built violent destruction recognition model, and the violent destruction recognition model fuses the multimodal monitoring feature vector to obtain a fused feature vector. Based on the fused feature vector, it is determined whether there is violent destruction behavior directed at the shared electric motorcycle. If so, the category of the violent destruction behavior corresponding to the multimodal monitoring data is determined, thereby improving the accuracy of violent destruction behavior identification.

[0043] In some embodiments, the above method further comprises: Generate early warning information and issue alerts based on the type of violent and destructive behavior; Identify abnormal users who commit violent and destructive acts and mark them.

[0044] Optionally, the location information of the shared electric motorcycle is obtained, and an early warning message is generated based on the location information of the shared electric motorcycle and the category of the violent destructive behavior.

[0045] Optionally, after determining the category of the violent destructive behavior, the alarm device on the shared electric motorcycle is triggered to sound an alarm.

[0046] It is understandable that by detecting violent acts of destruction in real time and issuing alarms, potential safety hazards of shared electric motorcycles can be discovered and dealt with in a timely manner; by marking and controlling abnormal users, potential violent acts of destruction can be effectively blocked.

[0047] Figure 2 The flowchart of the training process of the violent destruction recognition model provided by the embodiment of the present invention is as follows. Figure 2 As shown, in some embodiments, the training process of the brute force damage recognition model includes: Step 210: Obtain sample multimodal monitoring data of a sample shared electric motorcycle; Step 220: preprocess the sample multimodal monitoring data to obtain preprocessed sample multimodal monitoring data; Step 230: extract features from the pre-processed sample multimodal monitoring data to obtain a sample multimodal monitoring feature vector; Step 240: Determine the category label of the violent destructive behavior corresponding to the sample multimodal monitoring data; Step 250: Using the sample multimodal monitoring feature vector as a training sample and the category label of the violent destruction behavior corresponding to the sample multimodal monitoring data as the sample label, an initial violent destruction recognition model is trained. After the training is completed, a violent destruction recognition model is obtained.

[0048] In some embodiments, the sample multimodal monitoring data includes sample image data, sample sound data and / or sample vibration data; the sample multimodal monitoring feature vector includes a sample image feature vector, a sample sound feature vector and / or a sample vibration feature vector.

[0049] Optionally, feature extraction is performed on the sample image data to obtain a sample image feature vector; feature extraction is performed on the sample sound data to obtain a sample sound feature vector; feature extraction is performed on the sample vibration data to obtain a sample vibration feature vector.

[0050] The following describes a system for identifying violent and destructive behaviors for shared electric motorcycles provided by an embodiment of the present invention. The system for identifying violent and destructive behaviors for shared electric motorcycles described below and the method for identifying violent and destructive behaviors for shared electric motorcycles described above can be referenced to each other.

[0051] Figure 3 This is a schematic diagram of the structure of a violent destructive behavior identification system for shared electric motorcycles provided by an embodiment of the present invention, as shown in FIG. Figure 3 As shown, the violent destructive behavior identification system 300 for shared electric motorcycles includes: An acquisition unit 310 is used to acquire multimodal monitoring data of a shared electric motorcycle; A preprocessing unit 320 is used to preprocess the multimodal monitoring data to obtain preprocessed multimodal monitoring data; A feature extraction unit 330 is used to extract features from the pre-processed multimodal monitoring data to obtain a multimodal monitoring feature vector; Identification unit 340 is configured to input the multimodal monitoring feature vector into a pre-built violent vandalism identification model. The violent vandalism identification model fuses the multimodal monitoring feature vector to obtain a fused feature vector. Based on the fused feature vector, it is determined whether there is violent vandalism targeting shared electric motorcycles. If so, the category of violent vandalism corresponding to the multimodal monitoring data is determined. Among them, the violent destruction recognition model is trained based on the sample multimodal monitoring feature vector corresponding to the sample multimodal monitoring data of the sample shared electric motorcycle.

[0052] Optionally, the multimodal monitoring data includes image data, sound data and / or vibration data; the multimodal monitoring feature vector includes an image feature vector, a sound feature vector and / or a vibration feature vector.

[0053] Optionally, the multimodal monitoring feature vectors are fused to obtain a fused feature vector, including: determining a weight of an image feature vector, a weight of a sound feature vector, and / or a weight of a vibration feature vector; The image feature vector, the sound feature vector and / or the vibration feature vector are fused according to the weight of the image feature vector, the weight of the sound feature vector and / or the weight of the vibration feature vector to obtain a fused feature vector.

[0054] Optionally, the violent destructive behavior identification system for shared electric motorcycles further includes: The early warning unit is used to generate early warning information and issue an alarm based on the type of violent and destructive behavior; The marking unit is used to identify abnormal users who commit violent and destructive behaviors and mark the abnormal users.

[0055] Optionally, the training process of the brute force attack recognition model includes: Obtain sample multimodal monitoring data of sample shared electric motorcycles; Preprocessing the sample multimodal monitoring data to obtain preprocessed sample multimodal monitoring data; Perform feature extraction on the preprocessed sample multimodal monitoring data to obtain the sample multimodal monitoring feature vector; Determine the category label of the violent destructive behavior corresponding to the sample multimodal monitoring data; The sample multimodal monitoring feature vector is used as a training sample, and the category label of the violent destruction behavior corresponding to the sample multimodal monitoring data is used as the sample label to train the initial violent destruction recognition model. After the training is completed, the violent destruction recognition model is obtained.

[0056] Optionally, the sample multimodal monitoring data includes sample image data, sample sound data and / or sample vibration data; the sample multimodal monitoring feature vector includes a sample image feature vector, a sample sound feature vector and / or a sample vibration feature vector.

[0057] It should be noted here that the violent destructive behavior identification system for shared electric motorcycles provided by the embodiment of the present invention can implement all the method steps implemented by the above-mentioned violent destructive behavior identification method embodiment for shared electric motorcycles, and can achieve the same technical effects. The parts and beneficial effects of this embodiment that are the same as those of the method embodiment will not be described in detail here.

[0058] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Figure 4 As shown, the electronic device may include: a processor (processor) 410 , a communication interface (Communications Interface) 420 , a memory (memory) 430 and a communication bus 440 , wherein the processor 410 , the communication interface 420 , and the memory 430 communicate with each other via the communication bus 440 . The processor 410 can call the logic instructions in the memory 430 to execute a method for identifying violent destructive behavior towards shared electric motorcycles, the method including: obtaining multimodal monitoring data of shared electric motorcycles; preprocessing the multimodal monitoring data to obtain preprocessed multimodal monitoring data; extracting features from the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; inputting the multimodal monitoring feature vector into a pre-built violent destruction recognition model, and having the violent destruction recognition model fuse the multimodal monitoring feature vector to obtain a fused feature vector; based on the fused feature vector, determining whether there is violent destructive behavior towards shared electric motorcycles, and if so, determining the category of the violent destructive behavior corresponding to the multimodal monitoring data; wherein the violent destruction recognition model is trained based on the sample multimodal monitoring feature vector corresponding to the sample multimodal monitoring data of the sample shared electric motorcycles.

[0059] Furthermore, the logic instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0060] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the violent destructive behavior identification method for shared electric motorcycles provided by the above methods, the method including: obtaining multimodal monitoring data of shared electric motorcycles; preprocessing the multimodal monitoring data to obtain preprocessed multimodal monitoring data; extracting features from the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; inputting the multimodal monitoring feature vector into a pre-constructed violent destructive behavior identification model, and the violent destructive behavior identification model fuses the multimodal monitoring feature vector to obtain a fused feature vector. According to the fused feature vector, it is judged whether there is violent destructive behavior towards shared electric motorcycles. If so, the category of violent destructive behavior corresponding to the multimodal monitoring data is determined; wherein, the violent destructive behavior identification model is trained based on the sample multimodal monitoring feature vector corresponding to the sample multimodal monitoring data of the sample shared electric motorcycle.

[0061] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for identifying violent destructive behavior towards shared electric motorcycles provided by the above-mentioned methods, the method comprising: obtaining multimodal monitoring data of shared electric motorcycles; preprocessing the multimodal monitoring data to obtain preprocessed multimodal monitoring data; extracting features from the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; inputting the multimodal monitoring feature vector into a pre-constructed violent destructive behavior recognition model, and having the violent destructive behavior recognition model fuse the multimodal monitoring feature vector to obtain a fused feature vector; judging whether there is violent destructive behavior towards shared electric motorcycles based on the fused feature vector, and if so, determining the category of the violent destructive behavior corresponding to the multimodal monitoring data; wherein the violent destructive behavior recognition model is trained based on sample multimodal monitoring feature vectors corresponding to sample multimodal monitoring data of sample shared electric motorcycles.

[0062] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0063] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for identifying violent destructive behavior on shared electric motorcycles, characterized in that: include: Obtain multimodal monitoring data of shared electric motorcycles; Preprocessing the multimodal monitoring data to obtain preprocessed multimodal monitoring data; Performing feature extraction on the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; Inputting the multimodal monitoring feature vector into a pre-built violent vandalism identification model, the violent vandalism identification model fuses the multimodal monitoring feature vector to obtain a fused feature vector. Based on the fused feature vector, it is determined whether there is violent vandalism directed at the shared electric motorcycle. If so, the category of violent vandalism corresponding to the multimodal monitoring data is determined; Among them, the violent destruction identification model is obtained by training based on the sample multimodal monitoring feature vector corresponding to the sample multimodal monitoring data of the sample shared electric motorcycle.

2. The method for identifying violent destructive behavior against shared electric motorcycles according to claim 1, characterized in that: The multimodal monitoring data includes image data, sound data and / or vibration data; the multimodal monitoring feature vector includes an image feature vector, a sound feature vector and / or a vibration feature vector.

3. The method for identifying violent destructive behavior against shared electric motorcycles according to claim 2, characterized in that: The multimodal monitoring feature vectors are fused to obtain a fused feature vector, including: determining a weight of the image feature vector, a weight of the sound feature vector, and / or a weight of the vibration feature vector; The image feature vector, the sound feature vector and / or the vibration feature vector are fused according to the weight of the image feature vector, the weight of the sound feature vector and / or the weight of the vibration feature vector to obtain the fused feature vector.

4. The method for identifying violent destructive behavior against shared electric motorcycles according to claim 1, characterized in that: The method further comprises: Generate early warning information and issue an alarm based on the category of the violent and destructive behavior; Determine the abnormal user who has committed the violent destructive behavior and mark the abnormal user.

5. The method for identifying violent destructive behavior against shared electric motorcycles according to claim 1, characterized in that: The training process of the brute force damage recognition model includes: Obtain sample multimodal monitoring data of sample shared electric motorcycles; Preprocessing the sample multimodal monitoring data to obtain preprocessed sample multimodal monitoring data; Performing feature extraction on the preprocessed sample multimodal monitoring data to obtain a sample multimodal monitoring feature vector; Determining a category label of the violent destructive behavior corresponding to the sample multimodal monitoring data; The sample multimodal monitoring feature vector is used as a training sample, and the category label of the violent destruction behavior corresponding to the sample multimodal monitoring data is used as a sample label to train an initial violent destruction recognition model. After the training is completed, the violent destruction recognition model is obtained.

6. The method for identifying violent destructive behavior against shared electric motorcycles according to claim 5, characterized in that: The sample multimodal monitoring data includes sample image data, sample sound data and / or sample vibration data; the sample multimodal monitoring feature vector includes a sample image feature vector, a sample sound feature vector and / or a sample vibration feature vector.

7. A violent destructive behavior identification system for shared electric motorcycles, characterized by: include: An acquisition unit, used to acquire multimodal monitoring data of shared electric motorcycles; a preprocessing unit, configured to preprocess the multimodal monitoring data to obtain preprocessed multimodal monitoring data; A feature extraction unit, configured to extract features from the preprocessed multimodal monitoring data to obtain a multimodal monitoring feature vector; an identification unit, configured to input the multimodal monitoring feature vector into a pre-built violent vandalism identification model, have the violent vandalism identification model fuse the multimodal monitoring feature vector to obtain a fused feature vector, and determine, based on the fused feature vector, whether there is violent vandalism directed at the shared electric motorcycle; if so, determine the category of violent vandalism corresponding to the multimodal monitoring data; Among them, the violent destruction identification model is obtained by training based on the sample multimodal monitoring feature vector corresponding to the sample multimodal monitoring data of the sample shared electric motorcycle.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, it implements the method for identifying violent destructive behavior for shared electric motorcycles as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the method for identifying violent destructive behavior for shared electric motorcycles as described in any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, it implements the method for identifying violent destructive behavior for shared electric motorcycles as described in any one of claims 1 to 6.

Images