Vehicle positioning system detection method and device

Abstract

The present application provides a vehicle positioning system detection method and device, which relates to the field of neural network technology. The method can quickly obtain the target feature vector corresponding to the vehicle to be detected by acquiring the target shooting video of the vehicle to be detected, and adopts the video frame image calculation method, and the obtained target feature vector can accurately characterize the running characteristics of the vehicle to be detected; the extracted target feature vector is input as input data into the trained positioning system detection model, so that the model can be used to identify the target feature vector, and the recognition result of the target feature vector with high accuracy can be obtained efficiently, and further according to the analysis of the recognition result of the target feature vector, the failure detection result of the positioning system of the vehicle to be detected can be obtained, because the efficiency and accuracy of the target feature vector recognition are improved by combining the model recognition method, thereby effectively improving the efficiency and accuracy of the positioning system failure detection.

Description

Technical Field

[0001] The present application relates to the field of neural network technology, and in particular to a vehicle positioning system detection method and device. Background Art

[0002] The development of positioning systems enables real-time monitoring of vehicle operation, thereby achieving precise vehicle operation control. However, when the Global Positioning System (GPS) on the vehicle fails, the vehicle operation status cannot be accurately detected. Therefore, positioning system failure detection becomes particularly important.

[0003] However, in the prior art, GPS failures caused by GPS anomalies or weak signals are mainly discovered manually, but this judgment method based on subjective consciousness results in low accuracy in GPS failure detection. Summary of the invention

[0004] The purpose of the present application is to provide a vehicle positioning system detection method and device in view of the above-mentioned deficiencies in the prior art, so as to solve the problems of low efficiency and poor accuracy of vehicle positioning system detection in the prior art.

[0005] To achieve the above purpose, the technical solution adopted in the embodiment of the present application is as follows:

[0006] In a first aspect, an embodiment of the present application provides a vehicle positioning system detection method, comprising:

[0007] Obtaining the current movement speed of the vehicle to be detected detected by the positioning system of the vehicle to be detected, and when the current movement speed meets a first preset condition, obtaining a target shot video corresponding to the vehicle to be detected, wherein the target shot video includes a running video of the vehicle to be detected shot within a preset time period before the current moment;

[0008] Determine each target feature vector corresponding to the target shot video according to information of images at multiple moments in the target shot video and information of a next frame of image corresponding to each moment;

[0009] Using a pre-trained positioning system detection model to identify each target feature vector corresponding to the target shot video, and determining a recognition result of each target feature vector, wherein the recognition result includes: background motion or background non-motion;

[0010] According to the recognition results of each target feature vector, it is determined whether the positioning system of the vehicle to be detected fails.

[0011] Optionally, determining each target feature vector corresponding to the target shot video according to information of images at multiple moments in the target shot video and information of a next frame of image corresponding to each moment includes:

[0012] Respectively determine information of an image at a first moment, information of an image at a second moment, and information of an image at a third moment in the target shot video;

[0013] Taking the image at the first moment, the image at the second moment, and the image at the third moment as the current frame image, respectively determining information of the next frame image corresponding to the first moment, information of the next frame image corresponding to the second moment, and information of the next frame image corresponding to the third moment;

[0014] Determine the target feature vector corresponding to the first moment according to the information of the image at the first moment and the information of the next frame image corresponding to the first moment, determine the target feature vector corresponding to the second moment according to the information of the image at the second moment and the information of the next frame image corresponding to the second moment, and determine the target feature vector corresponding to the third moment according to the information of the image at the third moment and the information of the next frame image corresponding to the third moment;

[0015] The target feature vector corresponding to the first moment, the target feature vector corresponding to the second moment, and the target feature vector corresponding to the third moment are used as target feature vectors corresponding to the target shot video.

[0016] Optionally, determining the target feature vector corresponding to the first moment according to the information of the image at the first moment and the information of the next frame of image corresponding to the first moment includes:

[0017] Performing frame difference calculation on information of the image at the first moment and information of a next frame image corresponding to the first moment and taking an absolute value thereof, to determine information of an intermediate frame image corresponding to the first moment;

[0018] Determining, according to the information of the intermediate frame image, a proportion of the number of elements in the intermediate frame image whose pixel values ​​meet a second preset condition;

[0019] The target feature vector corresponding to the first moment is determined according to the proportion of the number of elements whose pixel values ​​satisfy the second preset condition in the intermediate frame image.

[0020] Optionally, determining, according to the information of the intermediate frame image, a proportion of the number of elements whose pixel values ​​satisfy a second preset condition in the intermediate frame image includes:

[0021] Determine the proportion of the number of elements whose pixel values ​​in the intermediate frame image are the first data as the first vector element;

[0022] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the second data as the second vector element;

[0023] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the third data as the third vector element;

[0024] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the fourth data as a fourth vector element;

[0025] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the first data and less than or equal to the second data as a fifth vector element;

[0026] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the second data and less than or equal to the third data as a sixth vector element;

[0027] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the third data and less than or equal to the fourth data as the seventh vector element.

[0028] Optionally, determining the target feature vector corresponding to the first moment according to the proportion of the number of elements whose pixel values ​​satisfy a second preset condition in the intermediate frame image includes:

[0029] A target feature vector corresponding to the first moment is determined according to the first vector element, the second vector element, the third vector element, the fourth vector element, the fifth vector element, the sixth vector element, and the seventh vector element.

[0030] Optionally, the using a pre-trained positioning system detection model to identify each target feature vector corresponding to the target shot video and determining the recognition result of each target feature vector includes:

[0031] The pre-trained positioning system detection model is used to identify each target feature vector corresponding to the input target shot video, and the recognition result of each target feature vector is output, wherein the recognition result includes background motion, which is used to characterize that the vehicle to be detected is in a running state, or background non-motion, which is used to characterize that the vehicle to be detected is in a stationary state.

[0032] Optionally, determining whether the positioning system of the to-be-detected vehicle fails according to the recognition result of each target feature vector includes:

[0033] If the number of background motions in the recognition results of each target feature vector is greater than a preset threshold, it is determined that the positioning system of the vehicle to be detected has failed.

[0034] Optionally, the method further comprises:

[0035] Collecting multiple videos with moving backgrounds as positive samples and multiple videos with non-moving backgrounds as negative samples from videos shot by multiple sample vehicles;

[0036] Extract at least one feature vector combination of each positive sample to obtain a first training sample set, extract at least one feature vector combination of each negative sample to obtain a second training sample set, each first training sample in the first training sample set has label information: background motion, and each second training sample has label information: background non-motion;

[0037] The positioning system detection model is obtained by training using the first training sample set and the second training sample set.

[0038] Optionally, the adopting the first training sample set and the second training sample set to train and obtain the positioning system detection model includes:

[0039] Inputting each first training sample in the first training sample set into an initial detection model to obtain a prediction recognition result of each first training sample output by the initial detection model;

[0040] Inputting each second training sample in the second training sample set into the initial detection model to obtain a prediction recognition result of each second training sample output by the initial detection model;

[0041] Calculating a first loss function according to the predicted recognition results of each first training sample and the label information of each first training sample;

[0042] Calculating a second loss function according to the predicted recognition results of each second training sample and the label information of each second training sample;

[0043] Obtaining a target loss function according to the first loss function and the second loss function;

[0044] According to the target loss function and the preset loss function accuracy, the initial detection model is iteratively corrected to obtain the positioning system detection model.

[0045] In a second aspect, an embodiment of the present application further provides a vehicle positioning system detection device, including: an acquisition module, a determination module, an identification module, and an analysis module;

[0046] The acquisition module is used to acquire the movement speed of the vehicle to be detected at the current moment detected by the positioning system of the vehicle to be detected, and when the movement speed at the current moment meets the first preset condition, acquire the target shot video corresponding to the vehicle to be detected, and the target shot video includes the running video of the vehicle to be detected shot within a preset time period before the current moment;

[0047] The determination module is used to determine each target feature vector corresponding to the target shot video according to information of images at multiple moments in the target shot video and information of the next frame image corresponding to each moment;

[0048] The recognition module is used to recognize each target feature vector corresponding to the target shot video using a pre-trained positioning system detection model, and determine a recognition result of each target feature vector, wherein the recognition result includes: background motion or background non-motion;

[0049] The analysis module is used to determine whether the positioning system of the vehicle to be detected fails based on the recognition results of each target feature vector.

[0050] Optionally, the determination module is specifically used to respectively determine information of an image at a first moment, information of an image at a second moment, and information of an image at a third moment in the target shot video;

[0051] Taking the image at the first moment, the image at the second moment, and the image at the third moment as the current frame image, respectively determining information of the next frame image corresponding to the first moment, information of the next frame image corresponding to the second moment, and information of the next frame image corresponding to the third moment;

[0052] Determine the target feature vector corresponding to the first moment according to the information of the image at the first moment and the information of the next frame image corresponding to the first moment, determine the target feature vector corresponding to the second moment according to the information of the image at the second moment and the information of the next frame image corresponding to the second moment, and determine the target feature vector corresponding to the third moment according to the information of the image at the third moment and the information of the next frame image corresponding to the third moment;

[0053] The target feature vector corresponding to the first moment, the target feature vector corresponding to the second moment, and the target feature vector corresponding to the third moment are used as target feature vectors corresponding to the target shot video.

[0054] Optionally, the determination module is specifically configured to perform frame difference calculation on the information of the image at the first moment and the information of the next frame image corresponding to the first moment and take an absolute value to determine the information of the intermediate frame image corresponding to the first moment;

[0055] Determining, according to the information of the intermediate frame image, a proportion of the number of elements in the intermediate frame image whose pixel values ​​meet a second preset condition;

[0056] The target feature vector corresponding to the first moment is determined according to the proportion of the number of elements whose pixel values ​​satisfy the second preset condition in the intermediate frame image.

[0057] Optionally, the determination module is specifically used to determine the proportion of the number of elements whose pixel values ​​in the intermediate frame image are the first data as the first vector element;

[0058] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the second data as the second vector element;

[0059] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the third data as the third vector element;

[0060] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the fourth data as a fourth vector element;

[0061] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the first data and less than or equal to the second data as a fifth vector element;

[0062] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the second data and less than or equal to the third data as a sixth vector element;

[0063] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the third data and less than or equal to the fourth data as the seventh vector element.

[0064] Optionally, the determination module is specifically used to determine the target feature vector corresponding to the first moment based on the first vector element, the second vector element, the third vector element, the fourth vector element, the fifth vector element, the sixth vector element, and the seventh vector element.

[0065] Optionally, the determination module is specifically used to use the pre-trained positioning system detection model to identify each target feature vector corresponding to the input target shot video, and output the recognition result of each target feature vector, wherein the recognition result includes background motion, which is used to characterize that the vehicle to be detected is in a running state, or background non-motion, which is used to characterize that the vehicle to be detected is in a stationary state.

[0066] Optionally, the analysis module is specifically configured to determine that the positioning system of the vehicle to be detected has failed if the number of background motions in the recognition results of each target feature vector is greater than a preset threshold.

[0067] Optionally, the device further comprises a training module;

[0068] The training module is used to collect a plurality of videos with moving backgrounds as positive samples and a plurality of videos with non-moving backgrounds as negative samples from videos shot by a plurality of sample vehicles;

[0069] Extract at least one feature vector combination of each positive sample to obtain a first training sample set, extract at least one feature vector combination of each negative sample to obtain a second training sample set, each first training sample in the first training sample set has label information: background motion, and each second training sample has label information: background non-motion;

[0070] The positioning system detection model is obtained by training using the first training sample set and the second training sample set.

[0071] Optionally, the training module is specifically used to input each first training sample in the first training sample set into an initial detection model to obtain a prediction recognition result of each first training sample output by the initial detection model;

[0072] Inputting each second training sample in the second training sample set into the initial detection model to obtain a prediction recognition result of each second training sample output by the initial detection model;

[0073] Calculating a first loss function according to the predicted recognition results of each first training sample and the label information of each first training sample;

[0074] Calculating a second loss function according to the predicted recognition results of each second training sample and the label information of each second training sample;

[0075] Obtaining a target loss function according to the first loss function and the second loss function;

[0076] According to the target loss function and the preset loss function accuracy, the initial detection model is iteratively corrected to obtain the positioning system detection model.

[0077] In a third aspect, an embodiment of the present application provides a processing device, comprising: a processor, a storage medium and a bus, wherein the storage medium stores machine-readable instructions executable by the processor. When the processing device is running, the processor and the storage medium communicate through the bus, and the processor executes the machine-readable instructions to perform the steps of the method provided in the first aspect.

[0078] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method provided in the first aspect are executed.

[0079] The beneficial effects of this application are:

[0080] The present application provides a vehicle positioning system detection method and device. By acquiring a target shot video of a vehicle to be detected and adopting a video frame image calculation method, a target feature vector corresponding to the vehicle to be detected can be quickly obtained, and the obtained target feature vector can accurately characterize the operating characteristics of the vehicle to be detected; the extracted target feature vector is input as input data into a trained positioning system detection model, so that the model can be applied to identify the target feature vector, and an identification result of the target feature vector with high accuracy can be efficiently obtained. Further, based on the analysis of the identification result of the target feature vector, a failure detection result of the positioning system of the vehicle to be detected can be obtained. Since the efficiency and accuracy of the target feature vector identification are improved by combining the model identification method, the efficiency and accuracy of the positioning system failure detection are effectively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0082] Figure 1 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 1 ;

[0083] Figure 2 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 2 ;

[0084] Figure 3 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 3 ;

[0085] Figure 4 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 4 ;

[0086] Figure 5 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 5 ;

[0087] Figure 6 A schematic diagram of a vehicle positioning system detection device provided in an embodiment of the present application;

[0088] Figure 7A schematic diagram of the structure of a processing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0089] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.

[0090] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0091] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0092] Figure 1 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 1 The execution subject of this method can be a computer device or a cloud server, or of course, a vehicle-mounted central control device with sufficient processing power. Figure 1 As shown, the method may include:

[0093] S101. Obtain the current movement speed of the vehicle to be detected as detected by the positioning system of the vehicle to be detected, and when the current movement speed meets a first preset condition, obtain a target shot video corresponding to the vehicle to be detected, wherein the target shot video includes an operation video of the vehicle to be detected shot within a preset time period before the current moment.

[0094] First of all, it should be noted that the GPS (Global Positioning System) validity test results obtained by this method can be used for driver operation time efficiency analysis in business assessment scenarios. In this scenario, the vehicle speed detected by the GPS on the vehicle operated by the driver and the detected GPS failure results can be used to determine whether the driver has completed the work as required.

[0095] In addition, the GPS validity test result obtained by this method can also be used to feed back to the driver to remind the driver so that the driver can report the faulty GPS in time. Of course, the application scenario is not limited to the above.

[0096] Optionally, the vehicle to be detected refers to a vehicle equipped with a GPS device to be detected. When the GPS device to be detected is determined, the corresponding target shooting video can be obtained from the vehicle to be detected equipped with the GPS device to be detected.

[0097] The speed value of the vehicle to be detected can be read in real time, and here the vehicle speed value detected by the GPS on the vehicle to be detected can be read. When the speed value of the vehicle to be detected at the current moment is 0, the detection of the GPS of the vehicle to be detected can be triggered.

[0098] When the speed value of the vehicle at the current moment of detection is 0, it means that the GPS cannot detect the vehicle speed. Then, the GPS may be invalid at present. One possibility is that the GPS is faulty. It is also possible that the vehicle is currently traveling to a place with weak signal, the GPS signal is blocked, and the speed cannot be detected. At this time, triggering the start of the GPS detection of the vehicle to be detected can timely grasp the effectiveness of the GPS.

[0099] Optionally, after triggering the start of GPS detection of the vehicle to be detected, the target shot video corresponding to the vehicle to be detected can be obtained first, and the target shot video can be a video shot by the installed camera equipment during the operation of the vehicle to be detected. In a feasible method, the target shot video can be a video shot by the vehicle to be detected within a preset time period before the current moment.

[0100] For example, the target video may be a video 5-10 seconds before the current moment. Currently, the video 0-5 seconds before the current moment may also be obtained. In this case, the video at the current moment may also be a part of the target video.

[0101] Of course, there are no particularly strict requirements for the time period for selecting the target shooting video. In different application scenarios, the time period for selecting the target shooting video varies greatly.

[0102] When applied to the driver's work efficiency assessment scenario, since real-time performance is not a big concern, the driver's vehicle driving status can be reviewed after the driver completes the work. Therefore, the selected time period for the target shooting video can be any time period before the current moment.

[0103] When applied to driver-side reminders, since the driver is required to be reminded of the vehicle's GPS validity in real time, and the GPS validity is unlikely to change significantly in a short period of time, the selected time period of the target video should be as close to the current moment as possible to avoid the situation where the GPS validity at historical moments cannot be used to measure the GPS validity at the current moment when the selected time is far away from the current moment.

[0104] S102: Determine target feature vectors corresponding to the target shot video according to information of images at multiple moments in the target shot video and information of the next frame of image corresponding to each moment.

[0105] Optionally, multiple moments can be set from the target captured video, and the image at each moment can be taken as the current frame. Then, when the current frame is determined, the next frame of the current frame is also determined, so that the information of the images of the set multiple moments and the information of the next frame of the image corresponding to each moment can be obtained from the target captured video. Here, the image information can refer to the pixel information in the image.

[0106] Then, according to the information of the image at each moment and the information of the next frame image corresponding to each moment, each target feature vector corresponding to the target shot video can be determined. Among them, the number of target feature vectors is consistent with the number of set moments, and one moment can correspond to one target feature vector.

[0107] S103, using a pre-trained positioning system detection model to identify each target feature vector corresponding to the target shot video, and determining a recognition result of each target feature vector, where the recognition result includes: background motion or background non-motion.

[0108] Optionally, each target feature vector corresponding to the target shot video obtained above can be input as input data into a pre-trained positioning system detection model, and the detection result of the model obtained is the recognition result of each target feature vector, wherein the recognition result may include: background motion or background non-motion.

[0109] S104: Determine whether the positioning system of the vehicle to be detected fails based on the recognition results of each target feature vector.

[0110] Optionally, by combining the recognition results of each target feature vector, it can be analyzed whether the positioning system of the vehicle to be detected fails.

[0111] It should be noted that when the recognition result of the target feature vector is background motion, it can be explained that the vehicle to be detected was moving when the video corresponding to the time of shooting the target feature vector was taken, and when the recognition result of the target feature vector is background non-motion, it can be explained that the vehicle to be detected was stationary when the video corresponding to the time of shooting the target feature vector was taken.

[0112] Normally, if the GPS is not invalid and the vehicle is moving, the GPS can definitely detect the speed value. Then, combined with the above-mentioned detection trigger conditions, that is, the speed of the vehicle to be detected detected by the GPS is 0, the GPS cannot detect the vehicle's speed information, and the vehicle to be detected is moving, it can be determined that the GPS is invalid.

[0113] In summary, the vehicle positioning system detection method provided in the present embodiment can quickly obtain the target feature vector corresponding to the vehicle to be detected by acquiring the target shot video of the vehicle to be detected and adopting the video frame image calculation method, and the obtained target feature vector can accurately characterize the running characteristics of the vehicle to be detected; the extracted target feature vector is input as input data into the trained positioning system detection model, so that the model can be applied to identify the target feature vector, and the recognition result of the target feature vector with high accuracy can be efficiently obtained, and further according to the analysis of the recognition result of the target feature vector, the failure detection result of the positioning system of the vehicle to be detected can be obtained, because the efficiency and accuracy of the target feature vector recognition are improved by combining the model recognition method, thereby effectively improving the efficiency and accuracy of the positioning system failure detection.

[0114] Figure 2 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 2 Optionally, in step S102, determining each target feature vector corresponding to the target shot video according to information of images at multiple moments in the target shot video and information of the next frame image corresponding to each moment may include:

[0115] S201, respectively determining information of an image at a first moment, information of an image at a second moment, and information of an image at a third moment in a target shot video.

[0116] The first moment, the second moment, and the third moment here are not fixed moments. It is sufficient that the first moment, the second moment, and the third moment form an arithmetic progression, that is, the second moment is separated from the first moment and the third moment by the same time. Assuming that the first moment is the first second of the target video, the second moment can be the third second, and the third moment can be the fifth second.

[0117] S202, taking the image at the first moment, the image at the second moment and the image at the third moment as the current frame image, and respectively determining the information of the next frame image corresponding to the first moment, the next frame image corresponding to the second moment and the next frame image corresponding to the third moment.

[0118] Optionally, taking the first moment as an example, when the image at the first moment is taken as the current frame image, the next frame image of the current frame image can be taken as the next frame image corresponding to the first moment. Similarly, when the image at the second moment is taken as the current frame image, the next frame image of the current frame image can be taken as the next frame image corresponding to the second moment; when the image at the third moment is taken as the current frame image, the next frame image of the current frame image can be taken as the next frame image corresponding to the third moment.

[0119] S203. Determine the target feature vector corresponding to the first moment based on the information of the image at the first moment and the information of the next frame image corresponding to the first moment; determine the target feature vector corresponding to the second moment based on the information of the image at the second moment and the information of the next frame image corresponding to the second moment; determine the target feature vector corresponding to the third moment based on the information of the image at the third moment and the information of the next frame image corresponding to the third moment.

[0120] In some embodiments, the target feature vector corresponding to the first moment can be calculated based on the information of the image at the first moment and the information of the next frame image corresponding to the first moment. Similarly, the target feature vector corresponding to the second moment can be calculated based on the information of the image at the second moment and the information of the next frame image corresponding to the second moment; the target feature vector corresponding to the third moment can be calculated based on the information of the image at the third moment and the information of the next frame image corresponding to the third moment.

[0121] S204: Use the target feature vector corresponding to the first moment, the target feature vector corresponding to the second moment, and the target feature vector corresponding to the third moment as target feature vectors corresponding to the target shot video.

[0122] Optionally, the target feature vector corresponding to each moment in the target shot video obtained above is used as the target feature vector corresponding to the target shot video, that is, the number of target feature vectors corresponding to the target shot video is not fixed, and is consistent with the number of set moments.

[0123] In one achievable manner, in order to improve the credibility of the GPS failure detection result, at least three moments may be set in the target shot video, that is, at least three target feature vectors corresponding to the target shot video are calculated.

[0124] Figure 3 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 3 Optionally, in step S203, determining the target feature vector corresponding to the first moment according to the information of the image at the first moment and the information of the next frame of image corresponding to the first moment may include:

[0125] S301, performing frame difference calculation on information of an image at a first moment and information of a next frame image corresponding to the first moment and taking an absolute value to determine information of an intermediate frame image corresponding to the first moment.

[0126] In this embodiment, the first moment is taken as an example to illustrate the relevant calculations. The calculation of the target feature vectors corresponding to other moments is the same as that of the first moment, and can be performed with reference to the calculation of the first moment.

[0127] Assume that the information of the image at the first moment is cur_Frame, and the information of the next frame image corresponding to the first moment is next_Frame, then |next_Frame-cur_Frame| can be calculated, that is, the absolute value of the two frames of image is taken by the frame difference method to obtain the information of the intermediate frame image corresponding to the first moment, where cur_Frame can be the pixel value matrix of the image at the first moment, and next_Frame can be the pixel value matrix of the next frame image corresponding to the first moment. Then, the calculated information of the intermediate frame image corresponding to the first moment is also a pixel value matrix.

[0128] S302: Determine, according to information of the intermediate frame image, a proportion of elements in the intermediate frame image whose pixel values ​​meet a second preset condition.

[0129] Optionally, based on the information of the intermediate frame image obtained by the above calculation, the proportion of the number of elements in the intermediate frame image whose pixel values ​​meet the second preset condition can be determined respectively.

[0130] Of course, in practical applications, it is not necessary to calculate the proportion, but it is also possible to only determine the number of elements whose pixel values ​​meet the second preset condition, and calculate other parameters for characterizing the target feature vector based on the number of elements.

[0131] S303: Determine a target feature vector corresponding to the first moment according to a proportion of the number of elements whose pixel values ​​satisfy a second preset condition in the intermediate frame image.

[0132] Optionally, based on the proportion of the number of elements whose pixel values ​​satisfy the second preset condition in the obtained intermediate frame image, the target feature vector corresponding to the first moment can be determined.

[0133] Optionally, in step S302, determining the proportion of the number of elements in the intermediate frame image whose pixel values ​​meet the second preset condition according to the information of the intermediate frame image may include:

[0134] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are the first data as the first vector element; determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the second data as the second vector element; determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the third data as the third vector element; determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the fourth data as the fourth vector element; determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the first data and less than or equal to the second data as the fifth vector element; determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the second data and less than or equal to the third data as the sixth vector element; determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the third data and less than or equal to the fourth data as the seventh vector element.

[0135] In an achievable manner, according to the information of the intermediate frame image, the proportion of the number of elements with a pixel value equal to 0 in the intermediate frame image can be counted as percent_0, the proportion of the number of elements with a pixel value less than 10 as percent_10, the proportion of the number of elements with a pixel value less than 20 as percent_20, and the proportion of the number of elements with a pixel value less than 30 as percent_30. That is, the first data, the second data, the third data, and the fourth data can be 0, 10, 20, and 30, respectively. Of course, such a value is only a feasible value-taking method that can be implemented in this method, and is not limited to the above-mentioned values.

[0136] In addition, based on the determined percent_0, percent_10, percent_20, and percent_30, the proportion sub_10 of the number of elements in the intermediate frame image whose pixel values ​​are greater than 0 and less than or equal to 10, the proportion sub_20 of the number of elements in the intermediate frame image whose pixel values ​​are greater than 10 and less than or equal to 20, and the proportion sub_30 of the number of elements in the intermediate frame image whose pixel values ​​are greater than 20 and less than or equal to 30 can be further calculated.

[0137] Among them, sub_10=percent_10-percent_0; sub_20=percent_20-percent_10; sub_30=percent_30-percent_20.

[0138] Similarly, the calculation here is not limited to calculating the proportion of the number of elements with pixel values ​​greater than 0 and less than or equal to 10 (sub_10), the proportion of the number of elements with pixel values ​​greater than 10 and less than or equal to 20 (sub_20), and the proportion of the number of elements with pixel values ​​greater than 20 and less than or equal to 30 (sub_30). It can also be used to calculate the proportion of the number of elements with pixel values ​​in other intervals.

[0139] Optionally, in step S303, determining the target feature vector corresponding to the first moment according to the proportion of the number of elements whose pixel values ​​in the intermediate frame image meet the second preset condition may include: determining the target feature vector corresponding to the first moment according to the first vector element, the second vector element, the third vector element, the fourth vector element, the fifth vector element, the sixth vector element, and the seventh vector element.

[0140] Based on the above calculation, a first vector element percent_0, a second vector element percent_10, a third vector element percent_20, a fourth vector element percent_30, a fifth vector element sub_10, a sixth vector element sub_20, and a seventh vector element sub_30 may be obtained.

[0141] In one implementable manner, based on the first vector element, the second vector element, the third vector element, the fourth vector element, the fifth vector element, the sixth vector element, and the seventh vector element, the target feature vector corresponding to the first moment can be determined to be (percent_0, percent_10, percent_20, percent_30, sub_10, sub_20, sub_30).

[0142] Of course, in practical applications, the determined vector elements are not limited to: the first vector element, the second vector element, the third vector element, the fourth vector element, the fifth vector element, the sixth vector element, and the seventh vector element. Their number can be adjusted arbitrarily, and the target feature vector obtained by combining the determined vector elements is not limited to the form listed above, but can also have other transformed forms.

[0143] Similarly, the target feature vector corresponding to the second moment and the target feature vector corresponding to the third moment can be determined respectively.

[0144] Optionally, in step S103, a pre-trained positioning system detection model is used to identify each target feature vector corresponding to the target shot video, and the recognition result of each target feature vector is determined, which may include: using a pre-trained positioning system detection model to identify each target feature vector corresponding to the input target shot video, and outputting the recognition result of each target feature vector, wherein the recognition result includes background motion, which is used to characterize that the vehicle to be detected is in a running state, or background non-motion, which is used to characterize that the vehicle to be detected is in a stationary state.

[0145] Optionally, each target feature vector corresponding to the target shot video is input into the positioning system detection model, the target feature vector is identified by the positioning system detection model, and the identification result is output, wherein each target feature vector corresponds to a recognition result, and the recognition result may be background motion, and the recognition result of background motion can indicate that the vehicle to be detected is in a running state, and the recognition result may also be background non-motion, and the recognition result of background non-motion can indicate that the vehicle to be detected is in a stationary state.

[0146] Optionally, in step S104, determining whether the positioning system of the vehicle to be detected is invalid based on the recognition results of each target feature vector may include: if the number of background motions in the recognition results of each target feature vector is greater than a preset threshold, determining that the positioning system of the vehicle to be detected is invalid.

[0147] In one feasible manner, if the recognition results of the target feature vector corresponding to the first moment, the recognition results of the target feature vector corresponding to the second moment, and the recognition results of the target feature vector corresponding to the third moment have at least two recognition results corresponding to the target feature vector at the moment as background motion, then it can be determined that the positioning system of the vehicle to be detected has failed.

[0148] For example: the recognition result of the target feature vector corresponding to the first moment is background motion, the recognition result of the target feature vector corresponding to the second moment is background motion, and the recognition result of the target feature vector corresponding to the third moment is background non-motion. This can indicate that the vehicle to be detected is in motion when shooting the target video. Since the movement speed of the vehicle to be detected detected by the positioning system is 0, it means that the positioning system has failed.

[0149] Of course, the more the number of background motions corresponding to the target feature vectors at each moment are, the higher the credibility of the determined positioning system failure.

[0150] Next, the training method of the positioning system detection model is explained.

[0151] Figure 4 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 4 ; Optionally, the method of the present application may further include:

[0152] S401 . Collect a plurality of videos with moving backgrounds as positive samples and a plurality of videos with non-moving backgrounds as negative samples from videos shot by a plurality of sample vehicles.

[0153] First, a batch of sample vehicles can be selected, and videos shot by the sample vehicles can be extracted. The videos shot by the sample vehicles with moving backgrounds are used as positive samples, and the videos shot by the sample vehicles with non-moving backgrounds are used as negative samples.

[0154] In order to improve the training accuracy of the model, a larger number of positive and negative samples can be extracted.

[0155] S402, respectively extract at least one feature vector combination of each positive sample to obtain a first training sample set, respectively extract at least one feature vector combination of each negative sample to obtain a second training sample set, wherein each first training sample in the first training sample set has label information: background motion, and each second training sample has label information: background non-motion.

[0156] Optionally, a feature vector corresponding to a moment can be extracted from each positive sample as a first training sample, so that the first training samples are combined to obtain a first training sample set. Of course, feature vectors corresponding to multiple moments can also be extracted from each positive sample, that is, multiple feature vectors can be extracted from one positive sample, and multiple first training samples can be obtained accordingly, wherein the label information of each first training sample in the first training sample set is: background motion.

[0157] Similarly, a feature vector corresponding to a moment can be extracted from each negative sample as a second training sample, so as to obtain a second training sample set by combining the second training samples. The label information of each second training sample in the second training sample set is: background does not move.

[0158] S403: Use the first training sample set and the second training sample set to train and obtain a positioning system detection model.

[0159] Optionally, the first training sample set and the second training sample set mentioned above may be used as training samples to train a positioning system detection model.

[0160] Figure 5 Schematic diagram of the process of the vehicle positioning system detection method provided in the embodiment of the present application Figure 5 Optionally, in step S403, using the first training sample set and the second training sample set to train a positioning system detection model may include:

[0161] S501 , input each first training sample in the first training sample set into an initial detection model to obtain a prediction recognition result of each first training sample output by the initial detection model.

[0162] The model loss function calculation method used in this embodiment is a conventional calculation method.

[0163] For the first training sample set, it is input into the initial detection model, and the prediction recognition results of each first training sample output by the model can be obtained.

[0164] S502: Input each second training sample in the second training sample set into the initial detection model to obtain a prediction recognition result of each second training sample output by the initial detection model.

[0165] Similarly, for the second training sample set, it is input into the initial detection model, and the prediction recognition results of each second training sample output by the model can be obtained.

[0166] S503: Calculate a first loss function according to the predicted recognition results of each first training sample and the label information of each first training sample.

[0167] The loss function in this embodiment may adopt the MSE (Mean Square Error) function, and its calculation formula may be: Wherein, f(x) refers to the predicted recognition result of the first training sample, and y refers to the target recognition result (ie, label information) of the first training sample.

[0168] In this way, the first loss function can be calculated.

[0169] S504: Calculate a second loss function according to the predicted recognition results of each second training sample and the label information of each second training sample.

[0170] Similarly, continuing with the above calculation formula, the second loss function can be calculated based on the predicted recognition results of each second training sample and the label information of each second training sample.

[0171] S505: Obtain a target loss function according to the first loss function and the second loss function.

[0172] Then, the sum of the first loss function and the second loss function can be used as the target loss function.

[0173] S506: According to the target loss function and the preset loss function accuracy, the initial detection model is iteratively corrected to obtain a positioning system detection model.

[0174] Determine whether the target loss function meets the preset loss function accuracy. If so, stop correcting the initial detection model, use the network parameters of the current initial detection model as the network parameters of the positioning system detection model, and obtain the positioning system detection model.

[0175] If the target loss function does not meet the preset loss function accuracy, the network parameters of the initial detection model are further modified, and the recognition results of each first training sample in the first training sample set are predicted based on the modified initial detection model, as well as the recognition results of each second training sample in the second training sample set are predicted. A new target loss function is calculated based on the new prediction results and label information, and it is determined whether the new target loss function meets the preset loss function accuracy. The process is iterated until the new target loss function meets the preset loss function accuracy, and the modification is stopped to obtain the positioning system detection model.

[0176] In summary, the vehicle positioning system detection method provided in the present embodiment can quickly obtain the target feature vector corresponding to the vehicle to be detected by acquiring the target shot video of the vehicle to be detected and adopting the video frame image calculation method, and the obtained target feature vector can accurately characterize the running characteristics of the vehicle to be detected; the extracted target feature vector is input as input data into the trained positioning system detection model, so that the model can be applied to identify the target feature vector, and the recognition result of the target feature vector with high accuracy can be efficiently obtained, and further according to the analysis of the recognition result of the target feature vector, the failure detection result of the positioning system of the vehicle to be detected can be obtained, because the efficiency and accuracy of the target feature vector recognition are improved by combining the model recognition method, thereby effectively improving the efficiency and accuracy of the positioning system failure detection.

[0177] The following describes the device, equipment, storage medium, etc. used to execute the vehicle positioning system detection method provided by the present application. The specific implementation process and technical effects are described above and will not be repeated below.

[0178] Figure 6 The schematic diagram of a vehicle positioning system detection device provided in an embodiment of the present application, the functions implemented by the vehicle positioning system detection device correspond to the steps performed by the above method. The device can be understood as the above-mentioned computer device or cloud server, such as Figure 6 As shown, the device may include: an acquisition module 610, a determination module 620, an identification module 630, and an analysis module 640;

[0179] The acquisition module 610 is used to acquire the movement speed of the vehicle to be detected at the current moment detected by the positioning system of the vehicle to be detected, and when the movement speed at the current moment meets the first preset condition, acquire the target shot video corresponding to the vehicle to be detected, the target shot video includes the running video of the vehicle to be detected shot within a preset time period before the current moment;

[0180] A determination module 620, configured to determine each target feature vector corresponding to the target shot video according to information of images at multiple moments in the target shot video and information of a next frame of image corresponding to each moment;

[0181] The recognition module 630 is used to recognize each target feature vector corresponding to the target shot video using a pre-trained positioning system detection model, and determine the recognition result of each target feature vector, the recognition result including: background motion or background non-motion;

[0182] The analysis module 640 is used to determine whether the positioning system of the vehicle to be detected fails according to the recognition results of each target feature vector.

[0183] Optionally, the determination module 620 is specifically configured to respectively determine information of an image at a first moment, information of an image at a second moment, and information of an image at a third moment in the target shot video;

[0184] Taking the image at the first moment, the image at the second moment, and the image at the third moment as the current frame image, respectively determining information of the next frame image corresponding to the first moment, information of the next frame image corresponding to the second moment, and information of the next frame image corresponding to the third moment;

[0185] Determine the target feature vector corresponding to the first moment according to the information of the image at the first moment and the information of the next frame image corresponding to the first moment, determine the target feature vector corresponding to the second moment according to the information of the image at the second moment and the information of the next frame image corresponding to the second moment, and determine the target feature vector corresponding to the third moment according to the information of the image at the third moment and the information of the next frame image corresponding to the third moment;

[0186] The target feature vector corresponding to the first moment, the target feature vector corresponding to the second moment, and the target feature vector corresponding to the third moment are used as target feature vectors corresponding to the target shot video.

[0187] Optionally, the determination module 620 is specifically configured to perform frame difference calculation on the information of the image at the first moment and the information of the next frame image corresponding to the first moment and take an absolute value to determine the information of the intermediate frame image corresponding to the first moment;

[0188] Determining, according to the information of the intermediate frame image, a proportion of the number of elements in the intermediate frame image whose pixel values ​​meet a second preset condition;

[0189] The target feature vector corresponding to the first moment is determined according to the proportion of the number of elements whose pixel values ​​satisfy the second preset condition in the intermediate frame image.

[0190] Optionally, the determination module 620 is specifically configured to determine a proportion of the number of elements whose pixel values ​​in the intermediate frame image are the first data as the first vector element;

[0191] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the second data as the second vector element;

[0192] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the third data as the third vector element;

[0193] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the fourth data as the fourth vector element;

[0194] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the first data and less than or equal to the second data as the fifth vector element;

[0195] Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the second data and less than or equal to the third data as a sixth vector element;

[0196] The proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the third data and less than or equal to the fourth data is determined as the seventh vector element.

[0197] Optionally, the determination module 620 is specifically configured to determine a target feature vector corresponding to the first moment according to the first vector element, the second vector element, the third vector element, the fourth vector element, the fifth vector element, the sixth vector element, and the seventh vector element.

[0198] Optionally, the determination module 620 is specifically used to use a pre-trained positioning system detection model to identify each target feature vector corresponding to the input target shot video, and output the recognition result of each target feature vector, wherein the recognition result includes background motion, which is used to characterize that the vehicle to be detected is in a running state, or the background is not moving, which is used to characterize that the vehicle to be detected is in a stationary state.

[0199] Optionally, the analysis module 640 is specifically configured to determine that the positioning system of the vehicle to be detected has failed if the number of background motions in the recognition results of each target feature vector is greater than a preset threshold.

[0200] Optionally, the device further comprises a training module;

[0201] A training module, for collecting a plurality of videos with moving backgrounds as positive samples and a plurality of videos with no moving backgrounds as negative samples from videos shot by a plurality of sample vehicles;

[0202] Extract at least one feature vector combination of each positive sample to obtain a first training sample set, extract at least one feature vector combination of each negative sample to obtain a second training sample set, each first training sample in the first training sample set has label information: background motion, and each second training sample has label information: background non-motion;

[0203] The first training sample set and the second training sample set are used to train a positioning system detection model.

[0204] Optionally, the training module is specifically used to input each first training sample in the first training sample set into the initial detection model to obtain the predicted recognition result of each first training sample output by the initial detection model;

[0205] Inputting each second training sample in the second training sample set into the initial detection model to obtain a prediction recognition result of each second training sample output by the initial detection model;

[0206] Calculating a first loss function according to the predicted recognition results of each first training sample and the label information of each first training sample;

[0207] Calculating a second loss function according to the predicted recognition results of each second training sample and the label information of each second training sample;

[0208] According to the first loss function and the second loss function, a target loss function is obtained;

[0209] According to the target loss function and the preset loss function accuracy, the initial detection model is iteratively corrected to obtain the positioning system detection model.

[0210] The above-mentioned device is used to execute the method provided by the aforementioned embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0211] The above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more digital singnal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0212] The above modules can be connected or communicated with each other via a wired connection or a wireless connection. The wired connection may include a metal cable, an optical cable, a hybrid cable, etc., or any combination thereof. The wireless connection may include a connection in the form of a LAN, a WAN, Bluetooth, a ZigBee, or NFC, or any combination thereof. Two or more modules may be combined into a single module, and any one module may be divided into two or more units. Those skilled in the art will clearly understand that for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the method embodiment, and will not be repeated in this application.

[0213] Figure 7 A structural schematic diagram of a processing device provided in an embodiment of the present application, which may be a computing device with data processing capabilities.

[0214] The device may include: a processor 801 and a storage medium 802 .

[0215] The storage medium 802 is used to store programs, and the processor 801 calls the programs stored in the storage medium 802 to execute the above method embodiment. The specific implementation method and technical effect are similar and will not be repeated here.

[0216] The storage medium 802 stores program code, and when the program code is executed by the processor 801, the processor 801 executes various steps of the method according to various exemplary embodiments of the present application described in the above “Exemplary Method” section of this specification.

[0217] Processor 801 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor (DigitalSignal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, and can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present application can be directly embodied as a hardware processor to be executed, or a combination of hardware and software modules in the processor can be executed.

[0218] Storage medium 802, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules. Storage medium may include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type storage medium, random access storage medium (Random Access Memory, RAM), static random access storage medium (Static Random Access Memory, SRAM), programmable read-only storage medium (Programmable Read Only Memory, PROM), read-only storage medium (Read Only Memory, ROM), electrically erasable programmable read-only storage medium (Electrically Erasable Programmable Read-Only Memory, EEPROM), magnetic storage medium, disk, optical disk, etc. Storage medium is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The storage medium 802 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0219] Optionally, the present application also provides a program product, such as a computer-readable storage medium, including a program, which is used to execute the above method embodiment when executed by a processor.

[0220] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0221] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0222] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0223] The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a computer-readable storage medium. The above-mentioned software functional unit is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (English: processor) to perform some steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only storage medium (English: Read-Only Memory, referred to as: ROM), random access storage medium (English: Random Access Memory, referred to as: RAM), disk or optical disk and other media that can store program code.

Claims

1. A vehicle positioning system detection method, characterized in that: include: Obtain the current speed of the vehicle to be detected detected by the positioning system of the vehicle to be detected, and when the current speed satisfies a first preset condition, obtain a target shot video corresponding to the vehicle to be detected, wherein the target shot video includes a running video of the vehicle to be detected taken within a preset time period before the current moment; the first preset condition is that the speed value of the vehicle to be detected at the current moment is 0; Determine each target feature vector corresponding to the target shot video according to information of images at multiple moments in the target shot video and information of a next frame of image corresponding to each moment; Using a pre-trained positioning system detection model to identify each target feature vector corresponding to the target shot video, and determining a recognition result of each target feature vector, wherein the recognition result includes: background motion or background non-motion; Determining whether the positioning system of the vehicle to be detected fails according to the recognition results of each target feature vector; The determining, based on information of images at multiple moments in the target shot video and information of a next frame of image corresponding to each moment, each target feature vector corresponding to the target shot video comprises: Respectively determine information of an image at a first moment, information of an image at a second moment, and information of an image at a third moment in the target shot video; Taking the image at the first moment, the image at the second moment, and the image at the third moment as the current frame image, respectively determining information of the next frame image corresponding to the first moment, information of the next frame image corresponding to the second moment, and information of the next frame image corresponding to the third moment; Determine the target feature vector corresponding to the first moment according to the information of the image at the first moment and the information of the next frame image corresponding to the first moment, determine the target feature vector corresponding to the second moment according to the information of the image at the second moment and the information of the next frame image corresponding to the second moment, and determine the target feature vector corresponding to the third moment according to the information of the image at the third moment and the information of the next frame image corresponding to the third moment; The target feature vector corresponding to the first moment, the target feature vector corresponding to the second moment, and the target feature vector corresponding to the third moment are used as target feature vectors corresponding to the target shot video.

2. The method according to claim 1, characterized in that The determining, according to information of the image at the first moment and information of the next frame of image corresponding to the first moment, a target feature vector corresponding to the first moment includes: Performing frame difference calculation on information of the image at the first moment and information of a next frame image corresponding to the first moment and taking an absolute value thereof, to determine information of an intermediate frame image corresponding to the first moment; Determining, according to the information of the intermediate frame image, a proportion of the number of elements in the intermediate frame image whose pixel values ​​meet a second preset condition; The target feature vector corresponding to the first moment is determined according to the proportion of the number of elements whose pixel values ​​satisfy the second preset condition in the intermediate frame image.

3. The method according to claim 2, characterized in that The determining, according to the information of the intermediate frame image, a proportion of the number of elements in the intermediate frame image whose pixel values ​​meet a second preset condition includes: Determine the proportion of the number of elements whose pixel values ​​in the intermediate frame image are the first data as the first vector element; Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the second data as the second vector element; Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the third data as the third vector element; Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are less than the fourth data as a fourth vector element; Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the first data and less than or equal to the second data as a fifth vector element; Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the second data and less than or equal to the third data as a sixth vector element; Determine the proportion of the number of elements in the intermediate frame image whose pixel values ​​are greater than the third data and less than or equal to the fourth data as the seventh vector element.

4. The method according to claim 3, characterized in that The determining, according to the proportion of the number of elements whose pixel values ​​in the intermediate frame image meet the second preset condition, the target feature vector corresponding to the first moment includes: A target feature vector corresponding to the first moment is determined according to the first vector element, the second vector element, the third vector element, the fourth vector element, the fifth vector element, the sixth vector element, and the seventh vector element.

5. The method according to any one of claims 1 to 4, characterized in that: The method of using a pre-trained positioning system detection model to identify each target feature vector corresponding to the target shot video and determining the recognition result of each target feature vector includes: The pre-trained positioning system detection model is used to identify each target feature vector corresponding to the input target shot video, and the recognition result of each target feature vector is output, wherein the recognition result includes background motion, which is used to characterize that the vehicle to be detected is in a running state, or background non-motion, which is used to characterize that the vehicle to be detected is in a stationary state.

6. The method according to claim 5, characterized in that The step of determining whether the positioning system of the vehicle to be detected fails according to the recognition results of each target feature vector includes: If the number of background motions in the recognition results of each target feature vector is greater than a preset threshold, it is determined that the positioning system of the vehicle to be detected has failed.

7. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Collecting multiple videos with moving backgrounds as positive samples and multiple videos with non-moving backgrounds as negative samples from videos shot by multiple sample vehicles; Extract at least one feature vector combination of each positive sample to obtain a first training sample set, extract at least one feature vector combination of each negative sample to obtain a second training sample set, each first training sample in the first training sample set has label information: background motion, and each second training sample has label information: background non-motion; The positioning system detection model is obtained by training using the first training sample set and the second training sample set.

8. The method according to claim 7, characterized in that The step of using the first training sample set and the second training sample set to train and obtain the positioning system detection model includes: Inputting each first training sample in the first training sample set into an initial detection model to obtain a prediction recognition result of each first training sample output by the initial detection model; Inputting each second training sample in the second training sample set into the initial detection model to obtain a prediction recognition result of each second training sample output by the initial detection model; Calculating a first loss function according to the predicted recognition results of each first training sample and the label information of each first training sample; Calculating a second loss function according to the predicted recognition results of each second training sample and the label information of each second training sample; Obtaining a target loss function according to the first loss function and the second loss function; According to the target loss function and the preset loss function accuracy, the initial detection model is iteratively corrected to obtain the positioning system detection model.

9. A vehicle positioning system detection device, characterized in that: include: Acquisition module, determination module, identification module, analysis module; The acquisition module is used to acquire the movement speed of the vehicle to be detected at the current moment detected by the positioning system of the vehicle to be detected, and when the movement speed at the current moment meets a first preset condition, acquire a target shot video corresponding to the vehicle to be detected, wherein the target shot video includes a running video of the vehicle to be detected taken within a preset time period before the current moment; the first preset condition is that the speed value of the vehicle to be detected at the current moment is 0; The determination module is used to determine each target feature vector corresponding to the target shot video according to information of images at multiple moments in the target shot video and information of the next frame image corresponding to each moment; The recognition module is used to recognize each target feature vector corresponding to the target shot video using a pre-trained positioning system detection model, and determine a recognition result of each target feature vector, wherein the recognition result includes: background motion or background non-motion; The analysis module is used to determine whether the positioning system of the vehicle to be detected fails according to the recognition results of each target feature vector; The determination module is specifically used to respectively determine the information of the image at the first moment, the information of the image at the second moment, and the information of the image at the third moment in the target shot video; Taking the image at the first moment, the image at the second moment, and the image at the third moment as the current frame image, respectively determining information of the next frame image corresponding to the first moment, information of the next frame image corresponding to the second moment, and information of the next frame image corresponding to the third moment; Determine the target feature vector corresponding to the first moment according to the information of the image at the first moment and the information of the next frame image corresponding to the first moment, determine the target feature vector corresponding to the second moment according to the information of the image at the second moment and the information of the next frame image corresponding to the second moment, and determine the target feature vector corresponding to the third moment according to the information of the image at the third moment and the information of the next frame image corresponding to the third moment; The target feature vector corresponding to the first moment, the target feature vector corresponding to the second moment, and the target feature vector corresponding to the third moment are used as target feature vectors corresponding to the target shot video.

Images