Sentinel mode-based automobile fuel tank oil stealing detection method and device

Abstract

The application provides a kind of based on sentry mode's automobile oil tank oil stealing detection method and device, it is related to automobile management technical field.The method comprises: setting up the area to be detected around automobile oil tank, the area to be detected includes guarded area and alarm area;Obtain the pedestrian image of the area to be detected, and determine suspected oil stealing object according to preset condition;According to the residence time of suspected oil stealing object in the guarded area and / or alarm area, determine the true oil stealing object.The based on sentry mode's automobile oil tank oil stealing detection method and device provided by the application can efficiently detect the oil stealing object, ensure the accuracy of the detected oil stealing object, avoid fuel theft, improve the safety of vehicle parking, and the application does not need to improve the automobile oil tank, only needs hardware equipment and software application to be combined with use, can greatly reduce the manufacturing cost and maintenance cost, while improving detection accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobile management, and in particular to a method and device for detecting oil theft from an automobile tank based on a sentinel mode, an electronic device, a non-transitory computer-readable storage medium, and a computer program product. BACKGROUND

[0002] In recent years, oil prices have been rising, and the transportation industry, which relies on fuel to ensure normal operation, has been severely affected. Some unscrupulous individuals take advantage of the gap when the driver is resting or leaving to use tools to damage the automobile tank and steal fuel, causing the driver to suffer huge property losses.

[0003] To solve the above problems, the current solution is mainly to improve the automobile tank. One is to install a sensor in the tank to detect abnormal oil level drop or foreign matter intrusion, and the other is to set a special locking structure on the tank. These two solutions have their own advantages and disadvantages. The sensor is sensitive, but it is easy to wear out and lose its effectiveness. As the use time increases, the probability of sensor false positives increases greatly, and the cost of frequent replacement is high. As for the tank with a special locking structure, it is difficult to manufacture, and it is difficult for the driver or the staff to operate when refueling, which is inconvenient. In addition, unscrupulous individuals usually prepare complete tools before stealing fuel, and can also damage the locking structure through violent means. Therefore, the above two solutions have not been able to efficiently solve the problem of fuel theft from automobiles. SUMMARY

[0004] The present application provides a method and device for detecting oil theft from an automobile tank based on a sentinel mode to solve the problem that the current solution of improving only the automobile tank has not been able to efficiently solve the problem of fuel theft from automobiles.

[0005] The present application provides a method for detecting oil theft from an automobile tank based on a sentinel mode, comprising:

[0006] Setting a detection area around the automobile tank, the detection area including a secured area and an alarm area;

[0007] Obtaining a pedestrian image of the detection area and determining a suspected oil theft object according to a preset condition;

[0008] Determining a real oil theft object according to the time the suspected oil theft object stays in the secured area and / or the alarm area.

[0009] According to the method for detecting oil theft from an automobile tank based on a sentinel mode provided by the present application, the pedestrian image of the detection area is obtained, and a suspected oil theft object is determined according to a preset condition, comprising:

[0010] According to the pedestrian image of the to-be-detected area, a pedestrian feature vector is obtained.

[0011] According to the similarity of the pedestrian feature vectors of the several pedestrian images obtained within a preset period, a suspected oil stealing object is determined.

[0012] According to the present application, a vehicle oil tank oil stealing detection method based on a sentinel mode is provided.

[0013] According to the pedestrian image of the to-be-detected area, a pedestrian target frame image is obtained through a target detection layer of a pedestrian detection model.

[0014] According to the pedestrian target frame image, a pedestrian feature vector is obtained through a feature vector extraction layer of the pedestrian detection model.

[0015] According to the present application, a vehicle oil tank oil stealing detection method based on a sentinel mode is provided.

[0016] The similarity of the pedestrian feature vectors of the several pedestrian images obtained within a preset period is greater than or equal to a preset similarity threshold.

[0017] According to the present application, a vehicle oil tank oil stealing detection method based on a sentinel mode is provided.

[0018] The stay time of the suspected oil stealing object in the secured area exceeds a first preset time threshold and / or the stay time of the suspected oil stealing object in the alarm area exceeds a second preset time threshold.

[0019] According to the present application, a vehicle oil tank oil stealing detection method based on a sentinel mode is provided.

[0020] When the real oil stealing object is determined, an alarm is issued.

[0021] The present application also provides a vehicle oil tank oil stealing detection device based on a sentinel mode, comprising:

[0022] A to-be-detected area setting module is used to set a to-be-detected area around a vehicle oil tank, wherein the to-be-detected area comprises a secured area and an alarm area.

[0023] A suspected oil stealing object determination module is used to obtain a pedestrian image of the to-be-detected area and determine a suspected oil stealing object according to a preset condition.

[0024] A real oil stealing object determination module is used to determine a real oil stealing object according to the stay time of the suspected oil stealing object in the secured area and / or the alarm area.

[0025] The application further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for detecting oil stealing from a vehicle tank based on a sentinel mode according to any one of the above when executing the program.

[0026] The application further provides a non-transitory computer readable storage medium, which stores a computer program, wherein the computer program implements the steps of the method for detecting oil stealing from a vehicle tank based on a sentinel mode according to any one of the above when executed by a processor.

[0027] The application further provides a computer program product, which comprises a computer program, wherein the computer program implements the steps of the method for detecting oil stealing from a vehicle tank based on a sentinel mode according to any one of the above when executed by a processor.

[0028] The application provides a method and device for detecting oil stealing from a vehicle tank based on a sentinel mode, which can efficiently detect oil stealing objects and ensure the accuracy of the detected oil stealing objects by obtaining a pedestrian image of a detection area, determining a suspected oil stealing object according to a preset condition, and obtaining a real oil stealing object according to the residence time of the suspected oil stealing object in a guarded area and an alarm area, thereby avoiding fuel theft, improving the safety of vehicle parking, and greatly reducing the manufacturing and maintenance costs and improving the detection accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0030] Figure 1 is a flowchart of the method for detecting oil stealing from a vehicle tank based on a sentinel mode provided by the application;

[0031] Figure 2 is a structural schematic diagram of the device for detecting oil stealing from a vehicle tank based on a sentinel mode provided by the application;

[0032] Figure 3 is a structural schematic diagram of the electronic device provided by the application. DETAILED DESCRIPTION

[0033] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0034] The present application provides a kind of based on sentry mode's automobile oil tank oil stealing detection method, device and electronic equipment. Figures 1-3 The present application provides a kind of based on sentry mode's automobile oil tank oil stealing detection method, device and electronic equipment.

[0035] With reference to Figure 1 The present application provides a kind of based on sentry mode's automobile oil tank oil stealing detection method, which can include:

[0036] S110, the terminal side equipment sets up the area to be detected around the automobile oil tank, and the area to be detected includes the area to be defended and the alarm area.

[0037] S120, the pedestrian image of the area to be detected is acquired, and the suspected oil stealing object is determined according to the preset condition.

[0038] S130, according to the residence time of the suspected oil stealing object in the area to be defended and / or the alarm area, the real oil stealing object is determined.

[0039] It should be noted that the sentry mode refers to the demarcation of the area to be detected (recommended demarcation, the whole picture is the area to be defended by default in the case of not demarcation, and the picture is the alarm area in the middle), whether there is abnormal residence behavior in the video picture is determined by the way of deep learning image recognition, and the automobile oil tank oil stealing detection method based on the sentry mode can more accurately and efficiently detect the real oil stealing object.

[0040] It should be noted that the execution subject of the automobile oil tank oil stealing detection method based on the sentry mode provided by the present application can be any terminal side equipment, such as automobile oil tank management terminal and the like.

[0041] In step S110, the terminal side equipment sets up the area to be detected around the automobile oil tank, and the area to be detected includes the area to be defended and the alarm area.

[0042] It should be noted that the position parameters of the area to be detected can be directly issued to the terminal side equipment to set up the area to be detected; the area to be detected can also be manually set through the mobile phone app (application) in communication connection with the terminal side equipment, server or multimedia intelligent central control equipment.

[0043] For example, the terminal-side device opens the WiFi, the mobile phone connects to the WiFi, and then enters the calibration interface through the app, selects the camera channel to be calibrated, and the terminal-side device transmits the corresponding channel picture and the existing region frame position to the mobile phone. The user can add, delete or modify the region frame in the picture. In the multimedia intelligent central control device, the calibration mode is selected to directly modify the region frame in the video picture. The background server transmits the information of the region frame to the terminal-side device by configuring the parameters of the terminal-side device.

[0044] After setting the to-be-detected region, the to-be-detected region can be divided into a defense region and an alarm region. The defense region can be a large irregular polygon region around the edge of the car, and the alarm region can be a rectangular region close to the oil tank position. The defense region can be adjacent to the alarm region or can be on the periphery of the alarm region.

[0045] In step S120, the terminal-side device acquires the pedestrian image of the to-be-detected region and determines a suspected oil stealing object according to a preset condition.

[0046] It should be noted that the terminal-side device can acquire the pedestrian image of the to-be-detected region through an infrared camera. The infrared camera can be installed below the rearview mirror on the side of the oil tank to monitor the region picture on the left side of the car and collect the pedestrian image of the to-be-detected region.

[0047] It should be noted that the preset condition can be that the same to-be-detected object is photographed more than a preset threshold number of pedestrian images in the to-be-detected region, which indicates that the to-be-detected object continuously stays in the to-be-detected region and can be determined as a suspected oil stealing object. The terminal-side device first determines the to-be-detected object with abnormal behavior according to the pedestrian image of the to-be-detected region and determines it as a suspected oil stealing object, which can improve the efficiency of detecting the real oil stealing object.

[0048] In step S130, the terminal-side device determines a real oil stealing object according to the stay time of the suspected oil stealing object in the defense region and / or the alarm region.

[0049] It should be noted that the terminal-side device can obtain the actual time of the suspected oil stealing object staying in the defense region and the alarm region, respectively, according to the time point of the suspected oil stealing object entering the defense region, the time point of the suspected oil stealing object leaving the defense region, the time point of the suspected oil stealing object entering the alarm region, and the time of the suspected oil stealing object leaving the alarm region. Then, the abnormal behavior of the suspected oil stealing object is estimated. When the actual time of staying in the defense region and / or the actual time of staying in the alarm region exceeds a threshold value, it indicates that there is a suspicion of stealing oil, and it is determined as a real oil stealing object.

[0050] The application provides a vehicle oil tank oil stealing detection method based on a sentinel mode, which can efficiently detect oil stealing objects, ensure the accuracy of the detected oil stealing objects, avoid fuel theft, improve the safety of vehicle parking, and greatly reduce the manufacturing and maintenance costs while improving the detection accuracy.

[0051] In an embodiment, step S120 can include:

[0052] According to the pedestrian image of the to-be-detected area, a pedestrian feature vector is obtained.

[0053] According to the similarity of the pedestrian feature vectors of the plurality of pedestrian images obtained within the preset time period, a suspected oil stealing object is determined.

[0054] Specifically, the pedestrian feature vector is obtained according to the pedestrian image of the to-be-detected area, including:

[0055] According to the pedestrian image of the to-be-detected area, a pedestrian target frame image is obtained through a target detection layer of a pedestrian detection model.

[0056] According to the pedestrian target frame image, a pedestrian feature vector is obtained through a feature vector extraction layer of the pedestrian detection model.

[0057] It should be noted that the pedestrian detection model can be any deep learning model suitable for embedded devices and used for identifying pedestrian objects, such as a Yolov5+repvgg pedestrian detection model, where Yolov5 is an algorithm framework and repvgg is the backbone network of the model. The pedestrian detection model can be pre-trained based on pedestrian detection samples.

[0058] It should be noted that when the terminal-side device obtains the pedestrian image of the to-be-detected area, it can first normalize (resize) it to a size of 320*320, then detect it through the target detection layer of the pedestrian detection model to obtain a pedestrian target frame, and then cut the image within the pedestrian target frame on the pedestrian image to obtain a pedestrian target frame image. Then, the pedestrian target frame image is normalized (resized) to a size of 128*256, and then the pedestrian feature vector is extracted through the feature vector extraction layer of the pedestrian detection model. Through the pedestrian detection model, a pedestrian feature vector with high precision can be quickly obtained. At the same time, the time point of each pedestrian image can be recorded as a basis for time determination.

[0059] It should be noted that before the pedestrian target frame image is obtained through the target detection layer of the pedestrian detection model, the fuzzy occlusion detection can be performed regularly to prevent human occlusion and camera smearing from causing experience problems and property losses to customers.

[0060] Specifically, a feature list can be set for each pedestrian image, and the feature list includes a pedestrian feature vector, an entering time point of the secured area, a leaving time point of the secured area, an entering time point of the alarm area, and a leaving time point of the alarm area. The entering time point refers to the time point of the first entering of the area, and the leaving time point refers to the time point of the last leaving of the area.

[0061] The several pedestrian images can be arranged according to time sorting according to the entering time point of the secured area, the leaving time point of the secured area, the entering time point of the alarm area, and the leaving time point of the alarm area, and the similarity of the pedestrian feature vectors of adjacent pedestrian images is calculated. If the similarity is less than a preset similarity threshold, it is considered to be different objects, and if the similarity is greater than or equal to the preset similarity threshold, it is considered to be the same object.

[0062] Alternatively, when the feature list is empty, a new object (such as a pedestrian feature vector) is inserted, and when the feature list is not empty, the similarity of the pedestrian feature vectors of all other pedestrian images and the newly transmitted pedestrian feature vector is compared, and the maximum similarity corresponding to the feature list is taken. The similarity is compared with a preset similarity threshold (for example, 0.6): if the similarity is less than 0.6, it is considered to be different pedestrians, and at this time the feature list can insert the next newly transmitted pedestrian feature; if the similarity is greater than or equal to 0.6, it is considered to be the same pedestrian.

[0063] It should be noted that the similarity can be calculated by the cosine distance, and the similarity threshold can also be adjusted according to specific circumstances.

[0064] If, within a preset time period, several pedestrian images are obtained, and there are several pedestrian images with a similarity exceeding a preset similarity threshold among the several pedestrian images, it indicates that a certain to-be-detected object in the to-be-detected area stays continuously within the preset time period, that is, the to-be-detected object can be determined as a suspected oil stealing object.

[0065] In an embodiment, the determination condition of the real oil stealing object is:

[0066] The stay time of the suspected oil stealing object in the secured area exceeds a first preset time threshold and / or the stay time of the suspected oil stealing object in the alarm area exceeds a second preset time threshold.

[0067] It should be noted that the residence time of each to-be-detected object in the secured area and the alarm area can be calculated according to the secured area entry time point, the secured area exit time point, the alarm area entry time point, and the alarm area exit time point in the feature list.

[0068] Specifically, when the residence time of the suspected oil stealing object in the secured area exceeds 30 seconds, the suspected oil stealing object can be determined as a real oil stealing object, or when the residence time of the suspected oil stealing object in the alarm area exceeds 5 seconds, the suspected oil stealing object can be determined as a real oil stealing object, and the like. The specific time threshold can be set according to actual conditions (for example, false positive rate or vehicle parking position, etc.).

[0069] In addition, the time threshold can also be set, and a frame number threshold can be set to take the frame number as the judgment basis to prevent time lag caused by process resource competition and further miss information.

[0070] In an embodiment, further comprising:

[0071] When the real oil stealing object is determined, an alarm is issued.

[0072] It should be noted that the alarm can be a voice alarm issued by a player, for example, reading "Please check the abnormal behavior near the fuel tank, pay attention to safety", or the alarm information can be sent to the multimedia intelligent center screen in the cab for the driver to check in time.

[0073] Further, the driver whether has handled the abnormality can be inferred through subsequent detection. If the real oil stealing object cannot be determined through several subsequent determinations, it indicates that the abnormal situation has been handled, and the alarm is stopped. If the real oil stealing object still exists in the subsequent determination, and the existence time exceeds a preset alarm time threshold (for example, 10 seconds), an abnormality reminding information can be sent to the supervision server to remind the relevant personnel to inform the driver to handle the abnormal situation in time.

[0074] Further, when the real oil stealing object is determined, the terminal side device can store the complete video from the first frame when the abnormal behavior person (real oil stealing object) is detected to the final exit of the to-be-detected area as the basis for later investigation and analysis, and the video can be played through the multimedia intelligent center screen. The driver can choose to get off to check according to the video displayed on the center screen if the driver is resting on the car. If the driver thinks it is a false alarm, the alarm can be manually stopped, or the video can be uploaded to the supervision server when the abnormality reminding information is issued when the abnormal situation is found to be not handled. The information recorded in the abnormality reminding information alarm information includes: the time (year, month, day, hour, minute, and second) information and the location (GPS longitude and latitude) information when the alarm is issued.

[0075] In addition, regarding installation of hardware devices that can be used in the method for detecting oil stealing from a vehicle tank based on a sentinel mode provided by the application, the background server can be placed in a user machine room, and an engineering instrument management system is deployed in the background server. The infrared camera can be installed below the left side mirror of the vehicle (compatible with the left side BSD (Blind Spot Vehicle Recognition System), except for special vehicle models), and the image contains the area to be detected. The terminal side device and the multimedia intelligent central control screen can be installed in the cab, and the installation position of the terminal side device and the multimedia intelligent central control screen is not too strict, as long as it is convenient for the driver to operate and watch, and does not hinder normal engineering operations. Considering the uncertainty of additional lighting conditions when parking at night outdoors, an infrared camera with infrared function is used, and when the lighting conditions are harsh, the camera automatically turns on the infrared fill light function.

[0076] When the terminal side device is powered on, the position information of the alarm area and the guarded area in the camera is read from the configuration file (the terminal side device will store the information in the configuration file after obtaining the alarm area and the guarded area by any means), and the terminal side device without obtaining the position information of the area will set the whole image as the guarded area, the middle of the image as the alarm area, and then input the whole image into the pedestrian detection model.

[0077] The following describes the device for detecting oil stealing from a vehicle tank based on a sentinel mode provided by the application, and the device for detecting oil stealing from a vehicle tank based on a sentinel mode described below can be mutually corresponding with the method for detecting oil stealing from a vehicle tank based on a sentinel mode described above.

[0078] Referring to Figure 2 The device for detecting oil stealing from a vehicle tank based on a sentinel mode provided by the application can include:

[0079] The area to be detected setting module 210 is configured to set an area to be detected around the vehicle tank, and the area to be detected includes a guarded area and an alarm area.

[0080] The suspected oil stealing object determination module 220 is configured to obtain a pedestrian image of the area to be detected, and determine a suspected oil stealing object according to a preset condition.

[0081] The real oil stealing object determination module 230 is configured to determine a real oil stealing object according to a stay time of the suspected oil stealing object in the guarded area and / or the alarm area.

[0082] In an embodiment, the suspected oil stealing object determination module 220 includes:

[0083] The pedestrian feature vector obtaining submodule is configured to obtain a pedestrian feature vector according to the pedestrian image of the area to be detected.

[0084] The first determination submodule is used to determine suspected oil theft targets based on the similarity of pedestrian feature vectors of several pedestrian images acquired within a preset time period.

[0085] In one embodiment, the pedestrian feature vector acquisition submodule includes:

[0086] The pedestrian target bounding box image acquisition submodule is used to: obtain the pedestrian target bounding box image based on the pedestrian image of the area to be detected through the target detection layer of the pedestrian detection model;

[0087] The pedestrian feature vector extraction submodule is used to: obtain pedestrian feature vectors from the pedestrian target box image through the feature vector extraction layer of the pedestrian detection model.

[0088] In one embodiment, the criteria for determining the suspected oil thief are as follows:

[0089] The similarity of the pedestrian feature vectors of several pedestrian images acquired within a preset time period is greater than or equal to a preset similarity threshold.

[0090] In one embodiment, the criteria for determining the actual oil thief are as follows:

[0091] The suspected oil thief stays in the protected area for more than a first preset time threshold and / or the suspected oil thief stays in the alarm area for more than a second preset time threshold.

[0092] In one embodiment, it further includes:

[0093] The alarm module is used to issue an alarm when the actual oil theft is identified.

[0094] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute a vehicle fuel tank theft detection method based on sentry mode, the method including:

[0095] A detection area is set up around the vehicle's fuel tank, and the detection area includes a protected area and an alarm area;

[0096] Acquire pedestrian images of the area to be detected, and determine suspected oil thieves according to preset conditions;

[0097] According to the stay time of the suspected oil stealing object in the guarded area and / or the alarm area, a real oil stealing object is determined.

[0098] In addition, the logical instructions in the memory 830 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0099] In another aspect, the present application also provides a computer program product, which includes a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program can be executed by a processor to enable a computer to execute the oil stealing detection method based on the sentinel mode for the automobile fuel tank provided by the above-mentioned methods, and the method includes:

[0100] An area to be detected is arranged around the automobile fuel tank, and the area to be detected includes a guarded area and an alarm area;

[0101] A pedestrian image of the area to be detected is acquired, and a suspected oil stealing object is determined according to a preset condition;

[0102] According to the stay time of the suspected oil stealing object in the guarded area and / or the alarm area, a real oil stealing object is determined.

[0103] In another aspect, the present application also provides a computer program product, which includes a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program can be executed by a processor to enable a computer to execute the oil stealing detection method based on the sentinel mode for the automobile fuel tank provided by the above-mentioned methods, and the method includes:

[0104] An area to be detected is arranged around the automobile fuel tank, and the area to be detected includes a guarded area and an alarm area;

[0105] A pedestrian image of the area to be detected is acquired, and a suspected oil stealing object is determined according to a preset condition;

[0106] According to the stay time of the suspected oil stealing object in the guarded area and / or the alarm area, a real oil stealing object is determined.

[0107] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0108] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software plus the necessary general hardware platform, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software products, and the computer software products can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and include a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0109] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for detecting siphoning of fuel from a vehicle tank based on a sentinel mode, characterized in that, The method comprises the following steps: An area to be detected is set around the oil tank of the automobile, and the area to be detected comprises a guarded area and an alarm area; the guarded area is an irregular polygon area around the edge of the automobile in the picture of the automobile camera; the alarm area is a rectangular area close to the position of the oil tank of the automobile in the picture of the automobile camera; A pedestrian image of the area to be detected is obtained, and a suspected oil stealing object is determined according to a preset condition; A real oil stealing object is determined according to the staying time of the suspected oil stealing object in the guarded area and / or the alarm area; The determination condition of the real oil stealing object is that the staying time of the suspected oil stealing object in the guarded area exceeds a first preset time threshold and / or the staying time of the suspected oil stealing object in the alarm area exceeds a second preset time threshold. The step of obtaining the pedestrian image of the area to be detected and determining the suspected oil stealing object according to the preset condition comprises the following steps: A pedestrian feature vector is obtained according to the pedestrian image of the area to be detected; The similarity of the pedestrian feature vectors of a plurality of pedestrian images obtained within a preset time period is obtained to determine the suspected oil stealing object; the way of determining the suspected oil stealing object is that: A feature list is set for each pedestrian image, and the feature list comprises a pedestrian feature vector, a time point of entering the guarded area, a time point of leaving the guarded area, a time point of entering the alarm area and a time point of leaving the alarm area; the time point of entering the area refers to the time point of entering the area for the first time, and the time point of leaving the area refers to the time point of leaving the area for the last time; when the feature list is empty, a new pedestrian feature vector is inserted; when the feature list is not empty, the similarity of the pedestrian feature vectors of all other pedestrian images and the newly transmitted pedestrian feature vector is compared, the maximum similarity corresponding feature list is taken, and the similarity is compared with a preset similarity threshold: if the similarity is less than the preset similarity threshold, it is considered that the pedestrians are different, and at this time, the feature list is inserted with the next newly transmitted pedestrian feature; if the similarity is greater than or equal to the preset similarity threshold, it is considered that the pedestrians are the same; The determination condition of the real oil stealing object is that the similarity of the pedestrian feature vectors of a plurality of pedestrian images obtained within a preset time period is greater than or equal to a preset similarity threshold. The step of obtaining the pedestrian feature vector according to the pedestrian image of the area to be detected comprises the following steps:

2. The method of claim 1, wherein the method further comprises: A pedestrian target frame image is obtained from a target detection layer of a pedestrian detection model according to the pedestrian image of the area to be detected; A pedestrian feature vector is obtained from a feature vector extraction layer of the pedestrian detection model according to the pedestrian target frame image. Further comprising:

3. The method of claim 1 or 2, wherein the method further comprises: When the real oil stealing object is determined, an alarm is sent out. The method comprises the following steps:

4. A device for detecting siphoning of fuel from a vehicle tank based on a sentinel mode, characterized in that An area to be detected is set around the oil tank of the automobile, and the area to be detected comprises a guarded area and an alarm area; the guarded area is an irregular polygon area around the edge of the automobile in the picture of the automobile camera; the alarm area is a rectangular area close to the position of the oil tank of the automobile in the picture of the automobile camera; ​ The suspected oil stealing object determination module is configured to: acquire a pedestrian image of the to-be-detected area, and determine a suspected oil stealing object according to a preset condition; and the acquisition of the pedestrian image of the to-be-detected area and the determination of the suspected oil stealing object according to the preset condition include: obtaining a pedestrian feature vector according to the pedestrian image of the to-be-detected area; and obtaining the determination of the suspected oil stealing object according to a similarity of the pedestrian feature vectors of a plurality of pedestrian images acquired within a preset time period; the determination of the suspected oil stealing object is performed in the following manner: a feature list is set for each pedestrian image, and the feature list includes: the pedestrian feature vector, an entering time point of the protected area, a leaving time point of the protected area, an entering time point of the alarm area, and a leaving time point of the alarm area, wherein the entering time point refers to a time point of first entering the area, and the leaving time point refers to a time point of last leaving the area; when the feature list is empty, a new pedestrian feature vector is inserted; when the feature list is not empty, a similarity of the pedestrian feature vectors of all other pedestrian images and a newly transmitted pedestrian feature vector is compared, a maximum similarity corresponding feature list is taken, and the similarity is compared with a preset similarity threshold value: if the similarity is less than the preset similarity threshold value, it is considered that the pedestrian is different, and at this time, the feature list is inserted with a next newly transmitted pedestrian feature; if the similarity is greater than or equal to the preset similarity threshold value, it is considered that the pedestrian is the same; and the determination condition of the suspected oil stealing object is that the similarity of the pedestrian feature vectors of the plurality of pedestrian images acquired within the preset time period is greater than or equal to the preset similarity threshold value. The real oil stealing object determination module is configured to: determine a real oil stealing object according to a staying time of the suspected oil stealing object in the protected area and / or the alarm area. The determination condition of the real oil stealing object is that: the staying time of the suspected oil stealing object in the protected area exceeds a first preset time threshold value and / or the staying time of the suspected oil stealing object in the alarm area exceeds a second preset time threshold value.

5. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the steps of the method for detecting automobile oil tank stealing based on the sentinel mode according to any one of claims 1 to 3. 6.A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method for detecting automobile oil tank stealing based on the sentinel mode according to any one of claims 1 to 3.

7. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method for detecting automobile oil tank stealing based on the sentinel mode according to any one of claims 1 to 3.

Images