Automobile blind area detection method and device, electronic equipment and storage medium

Abstract

The invention relates to an automobile blind area detection method and device, electronic equipment and a storage medium, and belongs to the technical field of automobile blind area detection.The method comprises the steps that blind area image data collected by an external camera of an automobile is obtained, feature extraction and monocular depth detection are conducted on the blind area image data, and depth information corresponding to each object can be obtained; further, the depth information of each object can be judged, and a judgment result is obtained; and then the blind area detection result of the automobile can be determined according to the judgment result, and the blind area detection result obtained according to the blind area image data acquired by the camera is not limited by the detection angle and is not interfered by radars of other automobiles, so that the detection accuracy is improved, and the risk of collision of the automobile is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile blind spot detection, and in particular to an automobile blind spot detection method, device, electronic equipment and storage medium. Background Art

[0002] Automobile blind spot detection is a relatively important function in the current automobile safety field. It can provide drivers with the distance between surrounding objects and themselves in real time, helping drivers reduce the possibility of traffic collision accidents. This technology mostly relies on sensors deployed on the car. Common blind spot detection technologies include radar sensors, camera systems, laser scanners, fusion sensor technology, and vehicle networking technology. Monocular depth detection is a relatively cutting-edge computer vision task. Its purpose is to infer the depth information in the real three-dimensional scene through the input two-dimensional image. It has important applications in augmented reality, robot navigation and other fields. At present, monocular depth detection is mainly achieved through rule-based methods and learning-based methods.

[0003] Currently, the most commonly used method for automobile blind spot detection is to use millimeter radar waves. Although this method has high reliability in most cases and can detect relatively accurate distances in real time, the cost of using millimeter radar waves is high, the detection angle is relatively limited, and it is easy to receive radar interference from other cars in dense traffic, resulting in detection.

[0004] Therefore, it is urgent to propose a vehicle blind spot detection method, device, electronic device and storage medium to solve the technical problems that the vehicle blind spot detection algorithm in the prior art has a limited detection angle and is easily interfered by radars from other vehicles in dense traffic, resulting in low detection accuracy and increasing the risk of vehicle collision. Summary of the invention

[0005] In view of this, it is necessary to provide a vehicle blind spot detection method, device, electronic device and storage medium to solve the technical problems existing in the prior art that the vehicle blind spot detection algorithm has a limited detection angle and is easily interfered by radars from other vehicles in dense traffic, resulting in low detection accuracy and increasing the risk of vehicle collision.

[0006] In order to solve the above problems, the present invention provides a method for detecting a blind spot of an automobile, comprising: Obtaining blind spot image data collected by an external camera of the vehicle; Performing feature extraction and monocular depth detection on the blind area image data to obtain depth information corresponding to each object; Determining the depth information of each object respectively to obtain a determination result; According to the judgment result, the blind spot detection result of the vehicle is determined.

[0007] In a possible implementation, performing feature extraction and monocular depth detection on the blind area image data to obtain depth information corresponding to each object includes: Preprocessing the blind area image data to obtain initial image data; Performing feature extraction on the initial image data to obtain multiple objects; Perform monocular depth detection on each object in the initial image data to obtain depth information corresponding to each object.

[0008] In a possible implementation manner, the respectively judging the depth information of each object to obtain a judgment result includes: Setting a collision risk level; the collision risk level includes a no risk level, a first risk level, and a second risk level; Determine whether the object is within a first preset range according to the depth information; the first preset range is the range of a first arc with a camera at a corresponding orientation as the center and a first preset distance as the radius; If not, determining that the judgment result of the object is the risk-free level; If so, a determination result is obtained according to the depth information of the object.

[0009] In a possible implementation manner, obtaining a determination result according to the depth information of the object includes: Determining whether the object is within a second preset range according to the depth information; the second preset range is a range of a second arc with the camera at the corresponding orientation as the center and the second preset distance as the radius; the second preset distance is smaller than the first preset distance; the first arc and the second arc are concentric arcs; If not, determining that the judgment result of the object is the first risk level; If so, it is determined that the judgment result of the object is the second risk level.

[0010] In a possible implementation, determining the blind spot detection result of the vehicle according to the judgment result includes: Performing color marking according to the collision risk level in the judgment result of each object to obtain color marking information; The blind spot detection result of the vehicle is determined according to the color marking information.

[0011] In a possible implementation, the first risk level is yellow marking information, the second risk level is red marking information, and determining the blind spot detection result of the vehicle according to the color marking information includes: When the color marking information includes the red marking information, determining that the blind spot detection result of the automobile is that there is a collision risk; When the color marking information includes the yellow marking information, determining that the blind spot detection result of the automobile is that there is a potential collision trend; When the color marking information does not include the red marking information and the yellow marking information, it is determined that the blind spot detection result of the automobile is no collision risk.

[0012] In a possible implementation, after determining the blind spot detection result of the vehicle according to the color labeling information, the method further includes: When the blind spot detection result indicates that there is a collision risk, determining a target object corresponding to the red marking information; Determining a recommended moving direction of the vehicle according to a line between the target object and a corresponding orientation camera; According to the suggested vehicle moving direction, prompt information is output.

[0013] On the other hand, the present invention also provides a vehicle blind spot detection device, comprising: An image acquisition module, used to acquire blind spot image data collected by an external camera of the vehicle; A depth calculation module is used to perform feature extraction and monocular depth detection on the blind area image data to obtain depth information corresponding to each object; An object determination module, used to determine the depth information of each object and obtain a determination result; The result determination module is used to determine the blind spot detection result of the vehicle according to the judgment result.

[0014] On the other hand, an embodiment of the present invention discloses an electronic device, including: a processor, a memory, and a computer program stored in the memory and capable of running on the processor, and when the computer program is executed by the processor, each step of the above-mentioned automobile blind spot detection method embodiment is implemented.

[0015] On the other hand, an embodiment of the present invention discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various steps of the above-mentioned embodiment of the vehicle blind spot detection method are implemented.

[0016] The beneficial effects of the present invention are as follows: by acquiring blind spot image data collected by an external camera of a car, and performing feature extraction and monocular depth detection on the blind spot image data, the depth information corresponding to each object can be obtained; then, the depth information of each object can be judged respectively to obtain a judgment result; and then, based on the judgment result, the blind spot detection result of the car can be determined, and the blind spot detection result obtained based on the blind spot image data collected by the camera will not be limited by the detection angle, nor will it be interfered by the radar of other cars, thereby improving the accuracy of detection and reducing the risk of car collision. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic flow chart of an embodiment of a vehicle blind spot detection method provided by the present invention; Figure 2 A schematic diagram of the structure of an embodiment of the vehicle blind distinction model provided by the present invention; Figure 3 For the present invention Figure 1 A schematic flow chart of an embodiment of step S102; Figure 4 For the present invention Figure 1 A schematic flow chart of an embodiment of step S103; Figure 5 A schematic diagram of the structure of an embodiment of the concentric arc model provided by the present invention; Figure 6 A schematic structural diagram of an embodiment of the vehicle blind spot detection device provided by the present invention; Figure 7 A schematic structural diagram of an embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0019] like Figure 1 As shown, a specific embodiment of the present invention discloses a method for detecting a blind spot of an automobile, comprising: S101, obtaining blind spot image data collected by an external camera of the vehicle; S102, performing feature extraction and monocular depth detection on the blind area image data to obtain depth information corresponding to each object; S103, judging the depth information of each object respectively to obtain a judgment result; S104: Determine the blind spot detection result of the vehicle according to the judgment result.

[0020] It should be understood that the blind spot of a car refers to the area that the driver cannot directly observe through the front windshield, side windows and rearview mirrors. Generally speaking, the blind spot of a car can be divided into 6 areas, namely the front blind spot, rear blind spot, left blind spot, right blind spot, A-pillar blind spot and B-pillar blind spot. The blind spot in front of the car is often easier to judge because it is observed from far to near. According to actual research, the blind spots of cars are mainly concentrated on the sides and rear of the car, which poses a greater collision safety hazard. Therefore, the left, right and rear areas of the car need to be paid special attention to. Therefore, consider installing the camera used to capture images for detection in these three locations, such as Figure 2 As shown, they are area 1, area 2 and area 3 respectively.

[0021] In a specific embodiment of the present invention, during the driving process of the car, the blind spot image data collected by the cameras installed on the left side, right side and rear area of ​​the car can be obtained in real time. The specific collection process and collection frequency can be set according to the actual situation, and the embodiment of the present invention is not limited here. Then, the blind spot image data can be subjected to feature extraction and monocular depth detection, and all objects existing in the blind spot image data and the depth information corresponding to each object can be obtained. Then, the depth information of each object can be judged respectively to obtain the judgment result of each object, and the blind spot detection result of the car can be determined based on the judgment results of all objects.

[0022] Compared with the prior art, the present embodiment provides a method for obtaining blind spot image data collected by an external camera of a car, performing feature extraction and monocular depth detection on the blind spot image data, and obtaining depth information corresponding to each object; then, the depth information of each object can be judged respectively to obtain a judgment result; then, based on the judgment result, the blind spot detection result of the car can be determined, and the blind spot detection result obtained based on the blind spot image data collected by the camera will not be limited by the detection angle, nor will it be interfered by the radar of other cars, thereby improving the accuracy of detection and reducing the risk of car collision.

[0023] It should be noted that the collected blind spot image data may have problems such as interference noise. In order to ensure the accuracy of detection, the blind spot image data needs to be processed.

[0024] In some embodiments of the present invention, Figure 3 As shown, step S102 includes: S301, preprocessing the blind area image data to obtain initial image data; S302, extracting features from the initial image data to obtain multiple objects; S303: Perform monocular depth detection on each object in the initial image data to obtain depth information corresponding to each object.

[0025] In a specific embodiment of the present invention, after obtaining the blind spot image data, the blind spot image data can be preprocessed, and the preprocessing can remove the interference noise in the image, and the denoised image can also be normalized to reduce the amount of system calculation, wherein the preprocessing process can include noise processing, normalization processing and other processes, and the specific preprocessing process can be set according to the actual situation, and the implementation of the present invention is not limited here. After preprocessing the blind spot image data, the initial image data can be obtained, and then the initial image data can be feature extracted, wherein the feature extraction is trained by deep learning to detect objects such as flat roads that exist in the blind spot but obviously do not cause collision risks to the car. The specific training process and the training model can be set according to the actual situation, and the embodiment of the present invention is not limited here. After obtaining all the objects in the blind spot image data, monocular depth detection can be performed on each object in the initial image data based on the monocular depth detection algorithm. The specific detection steps can be: the monocular depth detection is based on the Monodepth2 algorithm, and uses the trained model weight file resnet50_640x192. The initial image data is input into the monocular depth detection model, and a complete depth map is output. At the same time, the bounding box of each object in the image is detected using the target detection algorithm, and different images are segmented into different semantic areas, so as to distinguish different objects in the image. For each detected object, all depth values ​​in the object area are extracted from the depth map using its bounding box and segmentation mask; the average value of all depth values ​​in the object area is calculated as the overall depth distance of the object. Finally, the depth information corresponding to each object in the initial image data is obtained, that is, the distance of the object in the initial image data relative to the position of the shooting camera. Among them, the specific process and method of monocular depth detection can be set according to actual conditions, and the embodiment of the present invention is not limited here. After completing this step, the objects extracted by the aforementioned features are removed, and the objects in the remaining image and their corresponding depth information are processed in subsequent steps.

[0026] In some embodiments of the present invention, Figure 4 As shown, step S103 includes: S401, setting a collision risk level; the collision risk level includes a no risk level, a first risk level, and a second risk level; S402, determining whether the object is within a first preset range according to the depth information; the first preset range is a range of a first arc with a camera at a corresponding orientation as the center and a first preset distance as the radius; S403, if not, determining that the object is at a risk-free level; S404: If yes, obtain a judgment result based on the depth information of the object.

[0027] In a specific embodiment of the present invention, the collision risk level can be set according to the warning distance. The warning distance may include a first preset distance and a second preset distance. The collision risk level may include a no-risk level, a first risk level, and a second risk level, wherein the no-risk level indicates that there is no risk of collision, the first risk level indicates that there is a potential collision trend but no collision risk, and the second risk level indicates that there is a risk of collision. Then, it can be determined whether the corresponding object is within the first preset range according to the depth information of each object. If so, a judgment is made based on the depth information of the object to obtain a judgment result; if not, the judgment result of the object is determined to be a no-risk level, and it is determined that the object does not pose a collision risk to the car at this moment, and no processing is performed on the object. Among them, the first preset range can be the range of the first arc with the camera of the corresponding orientation as the center and the first preset distance as the radius.

[0028] In some embodiments of the present invention, step S404 includes: Determine whether the object is within a second preset range according to the depth information; the second preset range is the range of a second arc with the camera at the corresponding orientation as the center and the second preset distance as the radius; the second preset distance is smaller than the first preset distance; the first arc and the second arc are concentric arcs; If not, the judgment result of the object is determined to be the first risk level; If so, the judgment result of the object is determined to be the second risk level.

[0029] In a specific embodiment of the present invention, when an object is at a first preset distance, it is determined again based on the depth information of the object whether the object is in a second preset range. If so, this indicates that the object poses a risk of collision with the car, and is determined to be at the second risk level. If not, it is determined that the object has a potential collision tendency with the car at this moment, but does not have a collision risk, and the judgment result of the object is determined to be at the first risk level. When all objects are judged, all judgment results of all objects can be obtained. Among them, the second preset range is the range of the second arc with the camera of the corresponding orientation as the center and the second preset distance as the radius; Figure 5 As shown, the second preset distance R2 is smaller than the first preset distance R1; the first arc and the second arc are concentric arcs, C1+C2 is the range of the first arc of the first preset distance R1, and C2 is the range of the second arc of the second preset distance R2.

[0030] In some embodiments of the present invention, step S104 includes: Perform color marking according to the collision risk level in the judgment result of each object to obtain color marking information; Determine the blind spot detection result of the car based on the color marking information.

[0031] In a specific embodiment of the present invention, after obtaining the judgment results of all objects, the corresponding objects can be color-labeled according to the collision risk level in the judgment results of each object, and then the color labeling information corresponding to each object can be obtained; then the blind spot detection result of the car can be determined according to the color labeling information, for example, different levels in the judgment results can be labeled with different colors, and the specific labeled colors can be set according to the actual situation, and the embodiment of the present invention is not limited here.

[0032] In some embodiments of the present invention, the first risk level is yellow marking information, and the second risk level is red marking information. According to the color marking information, determining the blind spot detection result of the automobile includes: When the color marking information includes red marking information, it is determined that the blind spot detection result of the vehicle is that there is a collision risk; When the color marking information includes yellow marking information, it is determined that the blind spot detection result of the vehicle has a potential collision tendency; When the color marking information does not include red marking information and yellow marking information, it is determined that the blind spot detection result of the vehicle is no collision risk.

[0033] In a specific embodiment of the present invention, objects corresponding to the risk-free level may not be color-labeled, objects of the first risk level may be labeled yellow to obtain yellow labeling information, and objects of the second risk level may be labeled red to obtain red labeling information. Then, all color labeling information of all objects may be judged. When all color labeling information of all objects in the blind spot image data includes red labeling information, it indicates that there are objects with a collision risk within the second preset range of the car, and the zone detection result of the car may be determined to be a collision risk. When all color labeling information of all objects in the blind spot image data includes yellow labeling information, it indicates that there are objects with a potential collision trend but no collision risk outside the second preset range within the first preset range of the car, and the zone detection result of the car may be determined to be a potential collision trend but no collision risk. When all color labeling information of all objects in the blind spot image data does not include red labeling information and yellow labeling information, the blind spot detection result of the car is determined to be no collision risk.

[0034] In some embodiments of the present invention, after determining the blind spot detection result of the vehicle according to the color marking information, the method further includes: When the blind spot detection result indicates that there is a collision risk, determine the target object corresponding to the red marked information; Determine the recommended moving direction of the vehicle according to the connection line between the target object and the corresponding orientation camera; Output prompt information based on the suggested car moving direction.

[0035] In a specific embodiment of the present invention, after the judgment results of all objects are processed, the blind spot detection result can be obtained. When the blind spot detection result shows that there is a risk of collision, it means that there are target objects around the car that may collide. At this time, it is necessary to avoid the target objects in the blind spot of the car that may cause a collision risk to the car. The target object corresponding to the red marking information and the position of the shooting camera are connected as a straight line. The direction in which the target object points to the position of the shooting camera is the recommended direction of movement of the car. This direction is the direction away from the object with a risk of collision, for the driver's reference. Based on the above-mentioned system-recommended moving direction, in the processing part, when the car moves close to the target object, the system will issue a prompt sound to prompt that the car has a risk of collision. When the car moves away from the target object, it will re-judge in real time based on the distance after the car moves, completing a complete process of car blind spot detection, and providing reference and suggestions for the car to prevent collision.

[0036] The embodiment of the present invention first avoids the problem that conventional radar waves have too dense radar signals when facing dense vehicles, and may be interfered by signals from other vehicles. In terms of cost, the equipment of the embodiment of the present invention is simple, and only three cameras are needed to capture images for the improved barrage depth detection algorithm to detect, which greatly reduces the cost of use; in terms of performance, the embodiment of the present invention removes objects that will not cause collision risks through training before depth detection, which reduces the amount of calculation of the algorithm and is more suitable for real-time detection. At the same time, objects in the blind spot of the car are classified based on two distances, replacing the traditional gradient color of depth detection, which can more intuitively reflect the size of the collision risk caused by objects in the blind spot of the car to the car, and help the vehicle driving system and the driver to better avoid collision risks.

[0037] In order to better implement the automobile blind spot detection method in the embodiment of the present invention, based on the automobile blind spot detection method, the embodiment of the present invention also provides an automobile blind spot detection device, such as Figure 6 As shown, the automobile blind spot detection device 600 includes: An image acquisition module 601 is used to acquire blind spot image data collected by an external camera of the vehicle; The depth calculation module 602 is used to perform feature extraction and monocular depth detection on the blind area image data to obtain the depth information corresponding to each object; The object determination module 603 is used to determine the depth information of each object and obtain a determination result; The result determination module 604 is used to determine the blind spot detection result of the vehicle according to the judgment result.

[0038] The automobile blind spot detection device 600 provided in the above embodiment can implement the technical solution described in the above automobile blind spot detection method embodiment. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above automobile blind spot detection method embodiment, which will not be repeated here.

[0039] like Figure 7 As shown, the present invention also provides an electronic device 700. The electronic device 700 includes a processor 701, a memory 702 and a display 703. Figure 7 Only some components of the electronic device 700 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0040] In some embodiments, the memory 702 may be an internal storage unit of the electronic device 700, such as a hard disk or memory of the electronic device 700. In other embodiments, the memory 702 may also be an external storage device of the electronic device 700, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the electronic device 700.

[0041] Furthermore, the memory 702 may include both an internal storage unit of the electronic device 700 and an external storage device. The memory 702 is used to store application software installed in the electronic device 700 and various data.

[0042] In some embodiments, the processor 701 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 702, such as the automobile blind spot detection method of the present invention.

[0043] In some embodiments, the display 703 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 703 is used to display information of the electronic device 700 and to display a visual user interface. The components 701-703 of the electronic device 700 communicate with each other via a system bus.

[0044] In some embodiments of the present invention, when the processor 701 executes the automobile blind spot detection program in the memory 702, the following steps may be implemented: Obtaining blind spot image data collected by an external camera of the vehicle; Perform feature extraction and monocular depth detection on blind area image data to obtain the depth information corresponding to each object; Determine the depth information of each object and obtain the determination result; Based on the judgment result, the blind spot detection result of the car is determined.

[0045] It should be understood that: when the processor 701 executes the automobile blind spot detection program in the memory 702, in addition to the above functions, other functions can also be implemented. For details, please refer to the description of the corresponding method embodiment above.

[0046] Furthermore, the embodiment of the present invention does not specifically limit the type of the electronic device 700 mentioned, and the electronic device 700 may be a portable electronic device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, etc. Exemplary embodiments of portable electronic devices include but are not limited to portable electronic devices equipped with IOS, Android, Microsoft or other operating systems. The above-mentioned portable electronic devices may also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 700 may not be a portable electronic device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0047] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions of the vehicle blind spot detection method provided in the above-mentioned method embodiments can be implemented.

[0048] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0049] The above is a detailed introduction to the automobile blind spot detection method and device provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for detecting a vehicle blind spot, characterized in that: include: Obtaining blind spot image data collected by an external camera of the vehicle; Performing feature extraction and monocular depth detection on the blind area image data to obtain depth information corresponding to each object; Determining the depth information of each object respectively to obtain a determination result; According to the judgment result, the blind spot detection result of the vehicle is determined.

2. The vehicle blind spot detection method according to claim 1, characterized in that: The performing feature extraction and monocular depth detection on the blind area image data to obtain depth information corresponding to each object includes: Preprocessing the blind area image data to obtain initial image data; Performing feature extraction on the initial image data to obtain multiple objects; Perform monocular depth detection on each object in the initial image data to obtain depth information corresponding to each object.

3. The vehicle blind spot detection method according to claim 1, characterized in that: The determining the depth information of each object to obtain a determination result includes: Setting a collision risk level; the collision risk level includes a no risk level, a first risk level, and a second risk level; Determine whether the object is within a first preset range according to the depth information; the first preset range is the range of a first arc with a camera at a corresponding orientation as the center and a first preset distance as the radius; If not, determining that the judgment result of the object is the risk-free level; If so, a determination result is obtained based on the depth information of the object.

4. The vehicle blind spot detection method according to claim 3, characterized in that: The step of obtaining a judgment result according to the depth information of the object includes: Determining whether the object is within a second preset range according to the depth information; the second preset range is a range of a second arc with the camera at the corresponding orientation as the center and the second preset distance as the radius; the second preset distance is smaller than the first preset distance; the first arc and the second arc are concentric arcs; If not, determining that the judgment result of the object is the first risk level; If so, it is determined that the judgment result of the object is the second risk level.

5. The vehicle blind spot detection method according to claim 3, characterized in that: Determining the blind spot detection result of the vehicle according to the judgment result includes: Performing color marking according to the collision risk level in the judgment result of each object to obtain color marking information; The blind spot detection result of the vehicle is determined according to the color marking information.

6. The vehicle blind spot detection method according to claim 5, characterized in that: The first risk level is yellow marking information, the second risk level is red marking information, and determining the blind spot detection result of the vehicle according to the color marking information includes: When the color marking information includes the red marking information, determining that the blind spot detection result of the automobile is that there is a collision risk; When the color marking information includes the yellow marking information, determining that the blind spot detection result of the automobile is that there is a potential collision trend; When the color marking information does not include the red marking information and the yellow marking information, it is determined that the blind spot detection result of the automobile is no collision risk.

7. The vehicle blind spot detection method according to claim 6, characterized in that: After determining the blind spot detection result of the vehicle according to the color marking information, the method further includes: When the blind spot detection result indicates that there is a collision risk, determining a target object corresponding to the red marking information; Determining a recommended moving direction of the vehicle according to a line between the target object and a corresponding orientation camera; According to the suggested vehicle moving direction, prompt information is output.

8. A vehicle blind spot detection device, characterized in that: include: An image acquisition module, used to acquire blind spot image data collected by an external camera of the vehicle; A depth calculation module is used to perform feature extraction and monocular depth detection on the blind area image data to obtain depth information corresponding to each object; An object determination module, used to determine the depth information of each object and obtain a determination result; The result determination module is used to determine the blind spot detection result of the vehicle according to the judgment result.

9. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of the vehicle blind spot detection method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the vehicle blind spot detection method according to any one of claims 1 to 7 are implemented.

Images