Alarm method of detachable BSD alarm equipment, controller, medium and product

Abstract

The invention discloses a detachable BSD alarm device alarm method, a controller, a medium and a product, and relates to the field of blind area monitoring. When the method is implemented, the controller establishes a zero reference coordinate by taking a vehicle body fixed reference object as a reference and defines a blind area monitoring area, and can detect the position offset (including a translation offset value and a rotation offset angle) of the detachable camera in real time; when the position offset does not exceed the preset loosening threshold value, the blind area monitoring area is corrected through the coordinate compensation matrix, then the monitoring target is detected based on the corrected blind area monitoring area, and an alarm is given, so that detachable installation of the camera of the BSD alarm equipment is realized, the problem that the camera of the traditional fixed BSD alarm equipment is inconvenient to maintain and replace is solved, and the service life of the BSD alarm equipment is prolonged. And the monitoring area can be dynamically corrected when the detachable camera slightly deviates, so that the conditions of dislocation, missing detection or false alarm of the monitoring area caused by deviation of the camera are avoided, and the accuracy and the stability of blind area monitoring are ensured.

Description

Technical Field

[0001] This application relates to the field of blind spot monitoring, and in particular to an alarm method, controller, medium, and product for a detachable BSD alarm device. Background Technology

[0002] With the continuous growth of car ownership and the increasing complexity of road traffic environments, vehicle driving safety has received widespread attention. Blind Spot Detection (BSD), as an important active safety assistance system, can monitor the blind spots on both sides and rear of the vehicle in real time. When other vehicles or obstacles are detected entering the blind spot, it will promptly warn the driver, effectively reducing the risk of collisions during lane changes and turns.

[0003] Most BSD (Browser Detection and Disturbance) alarm devices in related technologies employ a fixed installation method, permanently fixing the camera to a specific location on the vehicle body surface using bolts or adhesive. During BSD alarm system initialization, a calibration program determines the camera's field of view and monitoring area, storing the calibration results as fixed parameters. In subsequent operation, the BSD alarm system performs target detection and recognition on the camera's captured images based on these fixed monitoring area parameters, triggering an alarm when a moving target appears within the monitoring area.

[0004] However, fixed installation presents maintenance inconveniences in practical applications. When a camera needs to be replaced due to collision, aging, or other reasons, the entire device must be removed from the vehicle body by disassembling bolts or damaging the adhesive layer, making the repair and replacement process complex and time-consuming. Summary of the Invention

[0005] This application provides an alarm method, controller, medium, and product for detachable BSD alarm devices, which improves the alarm accuracy of detachable BSD alarm devices.

[0006] Firstly, this application provides an alarm method for a detachable BSD alarm device, applied to the controller of the detachable BSD alarm device. The detachable BSD alarm device also includes a fixed base and a detachable camera. The fixed base is disposed on the surface of the vehicle body, and the detachable camera is connected to the fixed base via a mechanical snap-fit ​​structure. The method includes: receiving an initial monitoring image transmitted by the detachable camera; identifying the edge contour line of a fixed reference object on the vehicle body in the initial monitoring image; recording the coordinate position of the edge contour line in the image coordinate system of the initial monitoring image as a zero-position reference coordinate; extending a preset distance towards the rear of the vehicle in the initial monitoring image according to a preset spatial geometric relationship based on the position of the zero-position reference coordinate in the image coordinate system to delineate a blind spot monitoring area; acquiring the current monitoring image transmitted by the detachable camera, and in the current... The real-time edge contour line of the fixed reference object on the vehicle body is extracted from the front monitoring screen. The coordinate position of the real-time edge contour line is compared with the zero-position reference coordinate to calculate the position offset of the detachable camera relative to the installation zero position. The position offset includes translational offset value and rotational offset angle. It is determined whether the position offset exceeds the preset loosening threshold. If the position offset does not exceed the preset loosening threshold, a coordinate compensation matrix is ​​constructed based on the position offset. The coordinate range of the blind spot monitoring area is corrected by reverse translation and reverse rotation using the coordinate compensation matrix to obtain the corrected blind spot monitoring area. The pixel change features between consecutive video frames in the corrected blind spot monitoring area are extracted. Based on the spatial distribution area and movement direction of the pixel change features, it is determined whether there is a monitoring target. If there is a monitoring target, a blind spot alarm signal is generated and output.

[0007] By adopting the above technical solution, the controller establishes zero-position reference coordinates and delineates the blind spot monitoring area based on a fixed reference object on the vehicle body. It can also detect the positional offset of the detachable camera (including translational offset value and rotational offset angle) in real time. When the positional offset does not exceed the preset loosening threshold, the blind spot monitoring area is corrected through a coordinate compensation matrix. Then, based on the corrected blind spot monitoring area, the monitoring target is detected and an alarm is triggered. This not only realizes the detachable installation of the camera in the BSD alarm device, solving the problem of inconvenient maintenance and replacement of cameras in traditional fixed BSD alarm devices, but also dynamically corrects the monitoring area when the detachable camera has a slight positional offset, avoiding the situation of misalignment, missed detection, or false alarm caused by camera offset, thus ensuring the accuracy and stability of blind spot monitoring. At the same time, by identifying the monitoring target through pixel change features, it can accurately perceive the oncoming vehicles, pedestrians, and other targets in the rear blind spot of the vehicle, and output alarm signals in a timely manner, effectively reducing the collision risk when the vehicle changes lanes or turns.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, before receiving the initial monitoring image transmitted by the detachable camera, identifying the edge contour line of the fixed reference object on the vehicle body in the initial monitoring image, and recording the coordinate position of the edge contour line in the image coordinate system of the initial monitoring image as the zero reference coordinate, the method further includes: monitoring the electrical connection status between the fixed base and the detachable camera; when a positioning level signal generated by the detachable camera connecting to the fixed base is detected, outputting a locking command to the electromagnetic locking mechanism in the fixed base to drive the electromagnetic locking mechanism to mechanically lock the detachable camera; and after confirming that the locking is completed, opening the wireless video data transmission channel with the detachable camera.

[0009] By adopting the above technical solution, pre-steps such as electrical connection monitoring, electromagnetic locking, and opening of the wireless video data transmission channel are added after the detachable camera is in place. The controller detects the in-place level signal and triggers the electromagnetic locking mechanism to mechanically lock the detachable camera, ensuring the initial fixed stability of the detachable camera after installation and preventing the camera from easily loosening or shifting during vehicle movement. At the same time, the wireless video data transmission channel is opened only after locking is completed, ensuring that video data transmission between the detachable camera and the controller is carried out under the premise of reliable hardware fixation. This avoids initial reference coordinate calibration errors caused by transmitting data before the camera is properly installed, laying a hardware foundation for the accurate establishment of the zero-position reference coordinate and the reasonable delineation of the blind spot monitoring area, and further improving the reliability of the entire BSD alarm device.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, after determining whether the position offset exceeds the preset loosening threshold, the method further includes: if the position offset exceeds the preset loosening threshold, increasing the drive current output to the electromagnetic locking mechanism to a preset high engagement value to increase the mechanical clamping force of the fixed base on the detachable camera, and simultaneously generating a device abnormality warning signal and sending it to the vehicle display terminal.

[0011] By adopting the above technical solution, for situations where the positional offset exceeds the preset loosening threshold, a follow-up processing step is added: the controller increases the drive current of the electromagnetic locking mechanism and sends an abnormality warning signal. The controller adjusts the drive current to a preset high pull-in value, which increases the mechanical clamping force of the fixed base on the detachable camera, attempting to tighten the loose detachable camera in real time, restoring the stability of the camera's installation position as much as possible, and reducing monitoring failures caused by severe camera loosening. At the same time, an abnormality warning signal is sent to the vehicle display terminal, which can promptly feed back the fault information of camera loosening to the driver, reminding the driver to check and maintain the BSD alarm device. This achieves proactive early warning of equipment failure, taking into account both real-time fault remediation and manual maintenance prompts during equipment operation. This allows the detachable BSD alarm device to have both self-adjustment capabilities and timely feedback of problems that cannot be repaired by itself to the user, ensuring long-term stable operation of the device.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, pixel change features between consecutive video frames within the corrected blind zone monitoring area are extracted. Based on the spatial distribution area and motion direction of the pixel change features, it is determined whether a monitoring target exists. The monitoring target includes at least oncoming vehicle targets and pedestrian targets. Specifically, this includes: performing inter-frame difference calculation on consecutive video frames within the corrected blind zone monitoring area to obtain a pixel change amplitude distribution map; performing binarization processing on the pixel change amplitude distribution map to generate a motion region mask; extracting connected components from the motion region mask and calculating the geometric feature parameters of each connected component, including the area and aspect ratio of the connected component; tracking the centroid position change of each connected component in consecutive video frames to calculate the motion vector of each connected component; and determining whether a monitoring target exists based on the geometric feature parameters, motion vectors, preset vehicle shape thresholds, and preset pedestrian shape thresholds.

[0013] By adopting the above technical solution, the controller extracts the motion region through inter-frame difference calculation and binarization processing, and calculates the geometric feature parameters and motion vectors of the connected domain. Combined with preset vehicle shape thresholds and preset pedestrian shape thresholds, it judges the monitoring target, making the target recognition process more logical and accurate. It can effectively distinguish between oncoming vehicle targets and pedestrian targets in the blind spot, avoiding the misjudgment problem caused by simply relying on pixel change features. At the same time, by tracking the centroid position of the connected domain and calculating the motion vector, it can accurately capture the movement direction of the target, ensuring that dangerous targets approaching the vehicle can be identified in a timely manner. This provides an accurate target recognition basis for subsequent graded alarms, improves the accuracy and pertinence of BSD alarm equipment in identifying monitoring targets, and further ensures the reliability of blind spot monitoring during vehicle movement.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, if a monitoring target exists, a blind spot alarm signal is generated and output, specifically including: if a vehicle target exists, a first-level alarm signal is generated and output, the first-level alarm signal being a high-frequency flashing or rapid honking signal; if a pedestrian target exists, a second-level alarm signal is generated and output, the second-level alarm signal being a low-frequency flashing or voice prompt signal.

[0015] By adopting the above technical solution, a first-level alarm signal with high-frequency flashing or rapid beeping is output for oncoming vehicle targets, and a second-level alarm signal with low-frequency flashing or voice prompts is output for pedestrian targets. The controller can issue differentiated alarm prompts according to the different levels of danger of the monitored targets. Oncoming vehicle targets are usually characterized by high speed and high collision risk, and the rapid first-level alarm can quickly attract the driver's attention. However, pedestrian targets have different movement characteristics, and the second-level alarm with low-frequency flashing or voice prompts can effectively remind the driver without causing operational interference due to excessive alarms. This achieves refined output of alarm signals, allowing the driver to quickly judge the type of blind spot hazard based on the alarm type, improving the practicality and effectiveness of the alarm prompts, and further reducing the collision risk caused by different types of blind spot hazards.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, based on the position of the zero-position reference coordinates in the image coordinate system, a preset distance is extended towards the rear of the vehicle in the initial monitoring image according to a preset spatial geometric relationship to delineate the blind spot monitoring area. Specifically, this includes: using a preset pixel-distance mapping ratio to convert the preset distance into vertical pixel offsets and horizontal pixel offsets in the image coordinate system; using the zero-position reference coordinates as the reference vertex, determining the coordinates of the far-end boundary in the vertical direction of the image based on the vertical pixel offset, and determining the coordinates of the outer boundary in the horizontal direction of the image based on the horizontal pixel offset; and forming the blind spot monitoring area by the reference vertex, the far-end boundary coordinates, and the outer boundary coordinates.

[0017] By adopting the above technical solution, the controller converts the preset distance in the actual physical space into the pixel offset in the image coordinate system through the pixel-distance mapping ratio. Then, it determines the coordinates of the far-end boundary based on the zero-position reference coordinates and encloses the blind spot monitoring area. This achieves a precise mapping between the blind spot range in the actual physical space and the blind spot monitoring area in the image coordinate system, allowing the defined blind spot monitoring area to match the actual rear blind spot of the vehicle. This avoids the problem of the blind spot monitoring area being too large or too small due to deviations in space and pixel conversion. At the same time, the method of enclosing the area with the reference vertex and clear far-end boundary coordinates makes the delineation of the blind spot monitoring area more standardized and accurate. This lays a clear and standard coordinate foundation for subsequent area correction after camera offset, ensuring the scientific nature and accuracy of the blind spot monitoring area delineation.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, a coordinate compensation matrix is ​​constructed based on the position offset, and the coordinate range of the blind zone monitoring area is corrected by reverse translation and reverse rotation using the coordinate compensation matrix to obtain the corrected blind zone monitoring area. Specifically, this includes: determining a rotation factor based on the rotation offset angle, determining a translation factor based on the translation offset value, and combining the rotation factor and the translation factor to generate a coordinate mapping matrix; extracting the preset node coordinates of the blind zone monitoring area; performing coordinate transformation calculations on the preset node coordinates using the coordinate mapping matrix to obtain the corrected node coordinates; and reconstructing the boundary of the blind zone monitoring area based on the corrected node coordinates to obtain the corrected blind zone monitoring area.

[0019] By adopting the above technical solution, the correction process for the blind spot monitoring area after camera offset has been refined. The controller constructs a coordinate mapping matrix by rotating the offset angle and translating the offset value, transforms the preset node coordinates of the blind spot monitoring area and reconstructs the area boundary, making the correction process for the blind spot monitoring area more targeted and accurate. The rotating offset angle and translating the offset value are converted into quantified coordinate transformation factors. Through precise calculation of the preset node coordinates, the boundary of the entire blind spot monitoring area is reconstructed, which can effectively offset the monitoring area misalignment caused by slight camera offset. This ensures that the corrected blind spot monitoring area always matches the actual rear blind spot of the vehicle, avoiding missed detections and false detections caused by inaccurate area correction. At the same time, the node coordinate transformation method simplifies the calculation process of area correction and improves correction efficiency, allowing the BSD alarm device to complete the blind spot monitoring area correction in real time and quickly during vehicle operation, ensuring the continuity and accuracy of blind spot monitoring.

[0020] In a second aspect, embodiments of this application provide a controller comprising: one or more processors and a memory; the memory is coupled to the one or more processors and is used to store computer program code, the computer program code including computer instructions, wherein the one or more processors invoke the computer instructions to cause the controller to perform the method as described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on a controller, cause the controller to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a controller, cause the controller to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the controller provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. By adopting the above technical solution, the controller establishes zero-position reference coordinates and delineates the blind spot monitoring area based on the fixed reference object of the vehicle body. It can also detect the positional offset of the detachable camera (including translational offset value and rotational offset angle) in real time. When the positional offset does not exceed the preset loosening threshold, the blind spot monitoring area is corrected by the coordinate compensation matrix. Then, based on the corrected blind spot monitoring area, the monitoring target is detected and an alarm is triggered. This not only realizes the detachable installation of the camera of the BSD alarm device, solving the problem of inconvenient maintenance and replacement of the camera of the traditional fixed BSD alarm device, but also dynamically corrects the monitoring area when the detachable camera has a slight positional offset, avoiding the situation of misalignment, missed detection or false alarm caused by camera offset, ensuring the accuracy and stability of blind spot monitoring. At the same time, by identifying the monitoring target through pixel change features, it can accurately perceive the oncoming vehicles, pedestrians and other targets in the rear blind spot of the vehicle, and output alarm signals in a timely manner, effectively reducing the collision risk when the vehicle changes lanes or turns.

[0026] 2. By adopting the above technical solution, pre-steps such as electrical connection monitoring, electromagnetic locking, and opening of the wireless video data transmission channel after the detachable camera is in place are added. The controller detects the in-place level signal and triggers the electromagnetic locking mechanism to mechanically lock the detachable camera, ensuring the initial fixed stability of the detachable camera after installation and preventing the camera from easily loosening or shifting during vehicle movement. At the same time, the wireless video data transmission channel is opened only after locking is completed, ensuring that video data transmission between the detachable camera and the controller is carried out under the premise of reliable hardware fixation. This avoids initial reference coordinate calibration errors caused by transmitting data before the camera is properly installed, laying a hardware foundation for the accurate establishment of the zero-position reference coordinate and the reasonable delineation of the blind spot monitoring area, and further improving the reliability of the entire BSD alarm device.

[0027] 3. By adopting the above technical solution, the controller extracts the motion region through inter-frame difference calculation and binarization processing, and calculates the geometric feature parameters and motion vectors of the connected domain. Combined with preset vehicle shape thresholds and preset pedestrian shape thresholds, it judges the monitoring target, making the target recognition process more logical and accurate. It can effectively distinguish between oncoming vehicle targets and pedestrian targets in the blind spot, avoiding misjudgment problems caused by simply relying on pixel change features. At the same time, by tracking the centroid position of the connected domain and calculating the motion vector, it can accurately capture the movement direction of the target, ensuring that dangerous targets approaching the vehicle can be identified in a timely manner. This provides an accurate target recognition basis for subsequent graded alarms, improves the accuracy and pertinence of BSD alarm equipment in identifying monitoring targets, and further ensures the reliability of blind spot monitoring during vehicle movement. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating an alarm method for a detachable BSD alarm device in an embodiment of this application.

[0029] Figure 2 This is another flowchart illustrating the alarm method for the detachable BSD alarm device in this application embodiment;

[0030] Figure 3 This is a schematic diagram of the physical device structure of the controller in an embodiment of this application. Detailed Implementation

[0031] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0032] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0033] The following describes the process of the method provided in this implementation. Please refer to [link / reference]. Figure 1 This is a flowchart illustrating an alarm method for a detachable BSD alarm device in an embodiment of this application.

[0034] S101. Receive the initial monitoring image transmitted by the detachable camera, identify the edge contour line of the fixed reference object of the vehicle body in the initial monitoring image, and record the coordinate position of the edge contour line in the image coordinate system of the initial monitoring image as the zero reference coordinate.

[0035] The initial monitoring image consists of one or more frames of visual imagery of the vehicle's rear environment transmitted from the detachable camera to the controller after the detachable camera has completed mechanical snap-fit ​​assembly with the fixed base, the electrical contacts have made reliable contact, and the electromagnetic locking mechanism has performed its locking action and reached a stable locking state. This serves as the raw visual data for the controller to perform zero-position reference calibration, used to standardize the camera's installation posture and the monitoring area reference. The vehicle body fixed reference object refers to a structural component of the vehicle body that is fixed in position after leaving the factory, has a clear geometric outline, and does not undergo significant deformation or relative displacement under vehicle vibration, temperature changes, and normal external environmental interference. It provides a stable reference benchmark for camera installation posture calibration and offset detection, including but not limited to: stamped edges of the vehicle fenders, door frame edges, rigid contours of rearview mirror bases, and edges of the vehicle's taillight housings, etc., which possess straight or regular curved characteristics. Edge contour lines are geometric boundary lines with significant contrast and strong continuity formed between the vehicle body fixed reference object and other areas of the vehicle body and the external environment. In the image, they appear as continuous lines with drastic changes in pixel grayscale or gradient, which can be stably identified using conventional edge detection algorithms. The image coordinate system is a two-dimensional Cartesian coordinate system established using the pixel array of the initial monitoring image. Typically, the top-left corner of the image is the origin, the X-axis points horizontally to the right, and the Y-axis points vertically downwards. The coordinate unit is pixels, and it is used to quantify the position of targets in the image. The zero-position reference coordinates are the baseline data obtained by the controller after representing the coordinates of the identified fixed reference line of the vehicle body. This includes, but is not limited to: the set of pixel coordinates of several feature points on the edge contour line, the least-squares fitted straight line parameters of the edge contour line, and the centerline coordinates of the edge contour line. These serve as a unified reference for subsequent blind spot monitoring area delineation, camera offset detection, and coordinate compensation.

[0036] In practice, the controller receives the first frame or multiple consecutive frames of rear-side vehicle images sent by the detachable camera after assembly and locking via a wireless video data transmission channel established with the detachable camera as the initial monitoring screen. The controller performs image preprocessing on the initial monitoring screen, including Gaussian denoising, grayscale conversion, histogram equalization, Canny edge detection, or Sobel edge enhancement, to strengthen the edge features of the fixed reference object on the vehicle body and reduce the interference of ambient light and vehicle body reflection on contour recognition. After preprocessing, the controller searches for and matches image regions in the initial monitoring screen that match the preset fixed reference object on the vehicle body based on a preset contour template matching algorithm or deep learning recognition model, thereby locating the edge contour line of the fixed reference object. The controller extracts the pixel coordinates of key feature points such as inflection points, endpoints, and center points on the edge contour line using a pixel coordinate extraction algorithm, or performs straight line / curve fitting on the edge contour line to obtain the corresponding fitting parameters. The controller stores the above feature point coordinates or fitting parameters uniformly as zero-position reference coordinates to complete the zero-position calibration of the detachable camera in the current installation posture, providing a reference basis for subsequent blind spot monitoring area delineation and camera offset detection.

[0037] S102. Based on the position of the zero reference coordinates in the image coordinate system, extend a preset distance towards the rear of the vehicle in the initial monitoring image according to the preset spatial geometric relationship to delineate the blind spot monitoring area.

[0038] Among them, the preset spatial geometric relationship is a set of rules pre-stored in the controller and pre-calibrated based on the actual physical space range of the vehicle's rear blind spot. It is used to define the spatial positional relationship between the zero-position reference coordinates in the image coordinate system and the blind spot monitoring area. This set of rules is pre-calculated and determined in combination with the vehicle's specific body dimensions (including body width and rear overhang length), the installation angle of the detachable camera (horizontal installation tilt angle and vertical installation tilt angle), the camera's field of view (horizontal field of view angle and vertical field of view angle), and the camera lens focal length. It clarifies the direction, angle, and boundary constraints of extending the monitoring area to the rear of the vehicle from the zero-position reference coordinate as the reference starting point, providing a clear logical basis for the accurate delineation of the blind spot monitoring area and ensuring that the delineated image pixel area corresponds one-to-one with the actual physical blind spot at the rear of the vehicle.

[0039] The rear direction of the vehicle is taken as the normal driving direction and specifically refers to the left and right rear areas directly behind the vehicle (i.e., the core area covered by the blind spot of the vehicle's rearview mirror). This direction is the core monitoring direction of the BSD (Blind Spot Detection) alarm device and is also the main direction where blind spots are prone to collision risks during vehicle operation (especially when changing lanes or reversing). Its range can be adjusted according to the vehicle type (sedan, SUV, truck, etc.) in the preset spatial geometry.

[0040] The preset distance is a configurable parameter set in the controller to the actual physical space monitoring threshold corresponding to the rear blind spot of the vehicle. It can be set according to the driving safety standards, blind spot monitoring requirements, and relevant industry specifications of different vehicles. Specifically, it includes two types: one is the horizontal monitoring distance perpendicular to the vehicle's driving direction (i.e., the monitoring range to the side of the vehicle, used to cover the area of ​​oncoming vehicles in adjacent lanes), and the other is the longitudinal monitoring distance along the vehicle's driving direction (i.e., the monitoring range behind the vehicle, used to cover the area of ​​oncoming vehicles and pedestrians at close range behind). For example, the horizontal monitoring distance is set to 2.5-3.5 meters, and the longitudinal monitoring distance is set to 4-6 meters. This parameter can be adjusted and corrected later through the vehicle's OBD interface or the equipment's matching terminal to ensure adaptation to the blind spot characteristics of different vehicle models.

[0041] The blind spot monitoring area is a closed pixel area defined by the controller in the image coordinate system of the initial monitoring screen based on the zero-position reference coordinates, preset spatial geometric relationships, and preset distances. This area corresponds precisely to the actual physical blind spot at the rear of the vehicle through a pixel-distance mapping relationship. It is the core effective area for subsequent target acquisition by the detachable camera and target detection and alarm judgment by the controller. The pixel image outside this area is a non-monitoring area. The controller will not perform pixel change analysis and target recognition on it to reduce the controller's data processing load, improve monitoring efficiency, and avoid triggering false alarms by irrelevant targets in the non-blind spot area.

[0042] The specific implementation steps are as follows:

[0043] The first step is for the controller to retrieve the preset spatial geometric relationship parameters and preset distance parameters (including horizontal and vertical monitoring distances) stored internally. At the same time, it calls the pre-stored pixel-distance mapping ratio model. This pixel-distance mapping ratio model is established in advance based on the inherent optical parameters (lens focal length, imaging chip size) and actual installation parameters (installation height, installation tilt angle) of the detachable camera through geometric optics calculations and actual vehicle calibration. It clarifies the actual physical length (unit: mm / pixel) corresponding to one pixel unit in the image coordinate system. Moreover, this model can be adaptively corrected according to the slight adjustment of the camera's installation posture to ensure the accuracy of the conversion between pixels and actual distances.

[0044] The second step involves the controller converting the horizontal and vertical physical monitoring distances from the preset distances into corresponding horizontal and vertical pixel offsets in the image coordinate system, based on the pixel-distance mapping ratio model. During the conversion, the controller combines the vehicle rear direction rules defined in the preset spatial geometry to determine the specific pixel offset direction extending from the zero reference coordinate to the left and right rear of the vehicle (i.e., the offset direction of the X and Y axes in the image coordinate system), ensuring that the offset direction is consistent with the actual rear blind spot direction of the vehicle.

[0045] The third step involves the controller using the zero-position reference coordinates (specifically, the set of contour feature points or the fitted center line in the zero-position reference coordinates) as the core reference benchmark. Based on the pixel offset direction and pixel offset amount determined above, the controller initially plans the pixel range corresponding to the actual blind spot at the rear of the vehicle in the image coordinate system. During the planning process, the controller will simultaneously retrieve the pre-stored camera field of view parameters to perform boundary verification on the initially planned pixel range, eliminating invalid pixel areas that exceed the camera field of view range, thus avoiding the problem of invalid or missed monitoring areas due to field of view limitations.

[0046] The fourth step involves the controller performing coordinate boundary quantization on the verified pixel range, extracting the specific pixel coordinates of the four vertices (or key nodes of the closed contour) of the pixel range in the image coordinate system, connecting these coordinates sequentially to form a closed pixel region, and storing and marking the coordinate range of this region to identify it as the only valid area for subsequent target monitoring, thus completing the delineation of the blind spot monitoring area. After delineation, the controller associates and stores the coordinate parameters of the blind spot monitoring area with the zero-position reference coordinates to provide a comparison benchmark for subsequent area correction after camera offset.

[0047] S103. Obtain the current monitoring image transmitted by the detachable camera, extract the real-time edge contour line of the fixed reference object of the vehicle body in the current monitoring image, compare and calculate the coordinate position of the real-time edge contour line with the zero reference coordinate, and obtain the position offset of the detachable camera relative to the installation zero position. The position offset includes translation offset value and rotation offset angle.

[0048] The current monitoring screen refers to the rear-side visual image of the vehicle continuously and in real time received and transmitted by the detachable camera after the blind spot monitoring area is defined in step S102. This image serves as the real-time effective visual data for the controller to detect the positional deviation of the detachable camera and perform subsequent target monitoring. The acquisition frequency is consistent with the inherent frame rate of the detachable camera (usually 15-30 frames / second) to ensure that the controller can capture minute changes in the camera's position in a timely manner, avoiding lag in deviation detection due to excessively low acquisition frequency. At the same time, it takes into account the data processing load of the controller and avoids resource waste caused by excessively high frame rates.

[0049] Real-time edge contour lines refer to the edge contour lines of the fixed reference body of the vehicle body extracted by the controller in each frame of the current monitoring image through contour recognition logic that is completely consistent with step S101. It is the actual pixel contour representation of the fixed reference body of the vehicle body under the current camera installation posture. Its contour features are essentially the same as the initial edge contour lines in step S101 (both are the contours of the same fixed reference body of the vehicle body). The only difference is that due to possible positional shift of the camera, its coordinate position and rotation angle in the image coordinate system are different from the contour lines in the zero-position reference coordinates. This difference is the core basis for the controller to calculate the camera position offset.

[0050] Position offset refers to the quantitative data of the position change of the detachable camera relative to the initial installation zero position after the initial installation is completed (S101 step zero position calibration). This change is caused by factors such as vibration during vehicle operation, road bumps, minor external collisions, and slight loosening of the electromagnetic locking mechanism. It is a precise value calculated by the controller through a coordinate comparison algorithm. It is used to determine whether the camera installation is loose and whether blind spot monitoring area correction is needed. It is a core parameter to ensure the accuracy of blind spot monitoring.

[0051] Translational offset is an important component of positional offset. Specifically, it refers to the quantified positional change of the detachable camera in a plane parallel to the vehicle surface. This change occurs in the horizontal direction (parallel or perpendicular to the vehicle's direction of travel) and the vertical direction (a slight translation perpendicular to the vehicle surface, corresponding to the offset in the X and Y axes of the image coordinate system). In the image coordinate system, it is directly represented by the pixel offset difference in the X and Y axes of the feature point coordinates of the real-time edge contour line relative to the corresponding feature point coordinates in the zero-position reference coordinates. After conversion by the pixel-distance mapping ratio model, the translational distance (unit: millimeters) in the actual physical space is obtained. It is a one-dimensional numerical parameter and can be divided into X-axis translational offset and Y-axis translational offset, corresponding to the horizontal and vertical offsets in the image coordinate system, respectively.

[0052] Rotation offset angle is another core component of position offset. Specifically, it refers to the quantified angle (unit: degrees) of the detachable camera's clockwise or counterclockwise rotation around its installation center point relative to the fixed base. In the image coordinate system, it is represented as the angle between the fitted straight line / curve of the real-time edge contour and the fitted straight line / curve of the corresponding contour in the zero-position reference coordinate. Clockwise rotation is recorded as a positive angle, and counterclockwise rotation as a negative angle (or vice versa, a uniform rule can be preset in the controller). It is an angle-type parameter used to reflect the rotational changes in the camera's installation posture, avoiding blind spot monitoring area shifts, target misses, or false alarms caused by rotation.

[0053] The specific implementation steps are as follows:

[0054] The first step is for the controller to maintain a continuous wireless video data transmission channel with the detachable camera, and to receive the rear visual images of the vehicle captured by the detachable camera in real time at a frequency consistent with the camera's frame rate (15-30 frames / second). Each successfully received frame is defined as a current monitoring frame and is temporarily buffered to ensure the continuity of subsequent processing.

[0055] The second step involves the controller performing the same image preprocessing and contour recognition process as in step S101 for each frame of the current monitoring image. First, preprocessing algorithms such as Gaussian denoising, grayscale conversion, histogram equalization, and Canny edge enhancement are called to reduce noise and enhance the current monitoring image, eliminating interference factors such as changes in ambient lighting, vehicle reflections, and shadows from road debris, ensuring the clarity of the edge contour of the fixed reference object on the vehicle body. Then, the same preset contour template matching algorithm or deep learning recognition model as in step S101 is called to accurately retrieve and match the fixed reference object on the vehicle body in the preprocessed current monitoring image, locating its edge contour line. Finally, a pixel coordinate extraction algorithm is used to extract key feature points such as inflection points, endpoints, and center points on the real-time edge contour line, and the specific pixel coordinates of each feature point in the image coordinate system are recorded to form a real-time contour coordinate set.

[0056] The third step involves the controller calling a preset coordinate comparison algorithm to perform a point-by-point precise comparison between the real-time contour coordinate set obtained in the second step and the zero-position reference coordinates (the coordinate set of similar feature points) stored in step S101. First, the pixel offset difference (real-time coordinate value minus zero-position reference coordinate value) of each corresponding feature point in the X-axis and Y-axis directions of the image coordinate system is calculated to obtain the pixel offset matrix. Then, the pixel-distance mapping ratio model used in step S102 is called to convert each pixel difference in the pixel offset matrix into a displacement value in the actual physical space. The average value (or weighted average value, with the weight set according to the stability of the feature points) of the physical displacement values ​​of all feature points is taken to obtain the translation offset value of the detachable camera, which corresponds to the actual translation distance in the X-axis and Y-axis directions, respectively.

[0057] The fourth step involves the controller synchronously calling the contour rotation angle calculation algorithm to compare the rotation angle between the real-time edge contour line and the contour line in the zero-position reference coordinates. First, straight line / curve fitting is performed on the real-time contour coordinate set and the zero-position reference coordinate set to obtain two fitted lines (or fitted curves). Then, the angle between the two fitted lines is calculated using the vector angle calculation method. This angle is the rotation offset angle of the detachable camera relative to the initial installation zero position, and the rotation direction (clockwise / counterclockwise) is marked according to preset rules. If the fitted curve is not a straight line, feature vectors on the contour line can be selected to calculate the angle between the feature vectors to ensure the accuracy of the rotation angle calculation.

[0058] In the fifth step, the controller integrates and stores the translational offset values ​​(X-axis and Y-axis) calculated in the third step and the rotational offset angle calculated in the fourth step to form the complete positional offset of the detachable camera relative to the installation zero position.

[0059] S104. Determine whether the position offset exceeds the preset loosening threshold;

[0060] The preset loosening threshold refers to the critical value of positional offset pre-stored in the controller to determine whether the detachable camera has become abnormally loose. This threshold is set in combination with the camera installation accuracy and blind spot monitoring error tolerance, and is a combined judgment standard of translational offset value and rotational offset angle. For example, the X / Y axis translational offset value does not exceed 5 pixels and the rotational offset angle does not exceed 1°. Positional offset refers to the change in position of the detachable camera relative to the initial installation zero position, calculated by the controller through coordinate comparison. It is used to represent the actual degree of looseness of the camera, including quantitative data of two dimensions: translational offset value and rotational offset angle. Translational offset value refers to the pixel quantization value of the camera's horizontal and vertical translation in a plane parallel to the vehicle body surface, used to represent the degree of looseness of the camera's linear displacement. Rotational offset angle refers to the angle value of the camera's rotation around the installation point, in degrees, used to represent the degree of looseness of the camera's angular deflection.

[0061] In practice, the controller first retrieves a preset loosening threshold from local storage. This threshold includes the critical values ​​for the X-axis and Y-axis of the translational offset and the critical value for the rotational offset angle. It is a combined judgment standard, and the threshold requirements of each dimension must be met simultaneously to determine that the position is not loose. Then, the controller compares the actual translational offset value calculated in step S103 with the critical values ​​for the X-axis and Y-axis, and compares the actual rotational offset angle with the critical value for the angle. During the comparison, a two-dimensional joint judgment rule is adopted. If the translational offset value exceeds the critical value on any axis, or the rotational offset angle exceeds the critical value, it is determined that the positional offset exceeds the preset loosening threshold. If the translational offset value does not exceed the critical value on any axis, and the rotational offset angle does not exceed the critical value, it is determined that the positional offset does not exceed the preset loosening threshold. Finally, the controller temporarily stores the judgment result and triggers the corresponding subsequent steps based on the judgment result.

[0062] S105. If the position offset does not exceed the preset loosening threshold, a coordinate compensation matrix is ​​constructed based on the position offset. The coordinate compensation matrix is ​​used to perform reverse translation and reverse rotation correction on the coordinate range of the blind zone monitoring area to obtain the corrected blind zone monitoring area.

[0063] The coordinate compensation matrix is ​​a two-dimensional transformation matrix constructed by the controller based on the position offset, used to transform the coordinates of the blind zone monitoring area. It is a combination matrix of rotation and translation factors, used to achieve reverse translation and rotation correction of the monitoring area. The rotation factor is a matrix parameter calculated by the controller based on the rotation offset angle, used to represent the rotation transformation relationship in the image coordinate system. It is the core component of the coordinate compensation matrix, and its value is determined by the trigonometric function value of the rotation angle. The translation factor is a numerical parameter calculated by the controller based on the translation offset value, used to represent the translation transformation relationship in the image coordinate system, used to compensate for the linear displacement offset of the camera. The preset node coordinates refer to the coordinates of the nodes pre-selected by the controller on the boundary of the blind zone monitoring area. The selected feature coordinates are used to characterize the overall outline of the monitoring area, such as the four corner points and the midpoint of the boundary of the monitoring area. The corrected node coordinates refer to the new coordinates obtained by transforming the preset node coordinates through the coordinate compensation matrix, which are used to reconstruct the boundary of the corrected blind zone monitoring area. The corrected blind zone monitoring area refers to the new pixel area obtained after the blind zone monitoring area is corrected by reverse translation and rotation. This area matches the actual shooting angle after the camera offset and is the effective area for subsequent target detection. Reverse translation refers to the coordinate translation operation opposite to the actual translation direction of the camera, which is used to offset the translation offset of the camera. Reverse rotation refers to the coordinate rotation operation opposite to the actual rotation direction of the camera, which is used to offset the rotation offset of the camera.

[0064] The specific implementation steps are as follows:

[0065] The first step is parameter extraction and preprocessing: The controller first retrieves the position offset data stored in step S103, and identifies the rotation offset angle θ (unit: degrees) and translation offset values ​​(Δx, Δy). If the translation offset value is the actual physical displacement value (unit: millimeters), the pixel-distance mapping ratio model used in step S102 is called to convert it into pixel offset values ​​in the image coordinate system, ensuring that the calculation unit of all parameters is consistent with pixels. At the same time, the marking rule of the rotation offset angle (clockwise / counterclockwise is positive) is confirmed to provide a basis for the calculation of the rotation factor.

[0066] The second step involves calculating the rotation and translation factors: The controller uses trigonometric formulas to convert the rotation offset angle θ into corresponding sine (sinθ) and cosine (cosθ) values. Based on preset angle marking rules, a rotation factor (two-dimensional rotation matrix) is constructed. If clockwise rotation is considered a positive angle, the rotation factor matrix is: [cosθ, -sinθ; sinθ, cosθ]; if counterclockwise rotation is considered a positive angle, the rotation factor matrix is: [cosθ, sinθ; -sinθ, cosθ]; subsequently, based on the translation offset value (Δx, Δy), the translation factor (-Δx, -Δy) is calculated, ensuring that the direction of the translation factor is opposite to the actual translation direction of the camera.

[0067] The third step is the construction of the coordinate compensation matrix: The controller combines the rotation factor (two-dimensional rotation matrix) and translation factor (-Δx, -Δy) calculated in the second step to construct a complete 2×3 two-dimensional coordinate compensation matrix. The standard form of this matrix is: [cosθ, -sinθ, -Δx; sinθ, cosθ, -Δy] (taking clockwise rotation as positive as an example). This matrix can simultaneously realize coordinate transformations of reverse rotation and reverse translation. By substituting any preset node coordinates (x, y) into this matrix, the corresponding corrected node coordinates (x', y') can be obtained through matrix operations. The operation formulas are: x'=x×cosθ-y×sinθ-Δx, y'=x×sinθ+y×cosθ-Δy, ensuring the accuracy and efficiency of the transformation calculation.

[0068] The fourth step is the extraction of preset node coordinates: The controller extracts the set of preset node coordinates corresponding to the original blind zone monitoring area defined in step S102. This set contains the coordinates of all feature points that are pre-selected and can completely represent the contour of the monitoring area. After extraction, the preset node coordinates are sorted (in clockwise or counterclockwise order of the monitoring area boundary) to ensure the correctness of the order when reconstructing the boundary in the future and to avoid the problem of contour confusion.

[0069] Step 5: Calculation of node coordinate transformation: The controller substitutes each preset node coordinate (x, y) into the coordinate compensation matrix constructed in step 3, and performs coordinate transformation calculation according to the preset calculation formula (x'=x×cosθ-y×sinθ-Δx, y'=x×sinθ+y×cosθ-Δy). First, the rotation transformation is completed (achieved through the rotation factor), and then the translation transformation is completed (achieved through the translation factor), to obtain the corrected node coordinates (x', y') corresponding to each preset node coordinate. After the calculation is completed, all the corrected node coordinates are stored in the corresponding order to form a set of corrected node coordinates.

[0070] Step 6: Reconstruction and storage of the corrected blind zone monitoring area: The controller connects all coordinate points in the corrected node coordinate set one by one in the order of extraction (clockwise or counterclockwise) to form a closed pixel area boundary. This closed area is the corrected blind zone monitoring area. Subsequently, the controller associates and stores the coordinate range, correction time, and corresponding position offset data of the corrected blind zone monitoring area, overwriting the coordinate parameters of the original blind zone monitoring area, and confirming that this area is the only valid area for subsequent target detection, thus completing the correction operation of the entire blind zone monitoring area.

[0071] Optionally, in general, a coordinate compensation matrix is ​​constructed based on the position offset. The coordinate range of the blind zone monitoring area is then corrected by reverse translation and reverse rotation using the coordinate compensation matrix. The corrected blind zone monitoring area can be obtained in the following ways, which are not limited here: determine the rotation factor based on the rotation offset angle, determine the translation factor based on the translation offset value, and combine the rotation factor and translation factor to generate a coordinate mapping matrix; extract the preset node coordinates of the blind zone monitoring area; perform coordinate transformation calculation on the preset node coordinates using the coordinate mapping matrix to obtain the corrected node coordinates; reconstruct the boundary of the blind zone monitoring area based on the corrected node coordinates to obtain the corrected blind zone monitoring area.

[0072] S106. Extract the pixel change features between consecutive video frames within the corrected blind zone monitoring area, and determine whether there is a monitoring target based on the spatial distribution area and motion direction of the pixel change features.

[0073] Among them, pixel change features refer to the changes in brightness, color, and position of pixels between consecutive video frames within the corrected blind spot monitoring area. These features characterize the presence of moving targets within the area, including two core dimensions: spatial distribution area and direction of motion. Spatial distribution area refers to the pixel area occupied by the pixel change features in the image coordinate system, used to distinguish the size of moving targets. Direction of motion refers to the displacement direction of the pixel change features in consecutive video frames, used to determine whether a target is approaching the vehicle's blind spot. Consecutive video frames refer to the video frames within the corrected blind spot monitoring area continuously received by the controller at a preset frame rate, typically two or three adjacent frames, forming the basis for inter-frame difference calculation. Monitoring targets refer to moving objects that enter the vehicle's rear blind spot and pose a threat to vehicle safety, including at least oncoming vehicles and pedestrians, and are the core detection objects of the BSD alarm device. Oncoming vehicle targets refer to motor vehicles entering the vehicle's rear blind spot, characterized by high speed, large size, and high collision risk. Pedestrian targets refer to pedestrians entering the vehicle's rear blind spot, characterized by flexible movement, small size, and ease of being overlooked.

[0074] Specifically, the controller first extracts continuous video frames from the corrected blind zone monitoring area, selecting two or three adjacent frames as detection samples and removing invalid pixel data outside the area. Then, it performs inter-frame difference calculation on the selected continuous video frames, calculating the difference in grayscale or RGB values ​​between frames pixel by pixel. Pixels with differences greater than a preset pixel change threshold are marked as changed pixels, generating a pixel change amplitude distribution map to visually display the location and amplitude of pixel changes within the area. Next, the pixel change amplitude distribution map is binarized, setting changed pixels to 1 and unchanged pixels to 0, generating a motion region mask to mask the unchanged background area and retain only the area of ​​suspected moving targets. Then, connected components are extracted from the motion region mask, grouping adjacent pixels marked as 1 into a single connected component, removing noisy connected components with an area smaller than a preset minimum threshold, and finally calculating the geometric features of each effective connected component. The parameters include the pixel area and aspect ratio of the connected components. Simultaneously, the controller performs centroid tracking on each connected component, calculating its centroid coordinate changes across consecutive video frames. Combining the frame interval time and pixel-distance mapping ratio, a motion vector is obtained for each connected component, containing both motion speed and direction. Finally, the controller retrieves pre-stored preset vehicle and pedestrian shape thresholds. The preset vehicle shape threshold corresponds to a larger area range and a specific aspect ratio range (e.g., 3:1 to 5:1), while the preset pedestrian shape threshold corresponds to a smaller area range and another set of aspect ratio ranges (e.g., 1:2 to 1:3). The controller matches the geometric feature parameters of the connected components with the thresholds, and simultaneously uses the motion vector to determine if the target is moving towards the vehicle. If a match is successful and the direction of movement is in the vehicle's blind spot, a corresponding monitoring target is identified; otherwise, no monitoring target is identified.

[0075] Optionally, under normal circumstances, the pixel change features between consecutive video frames within the corrected blind zone monitoring area are extracted. Based on the spatial distribution area and motion direction of the pixel change features, it is determined whether a monitoring target exists. The monitoring targets include at least oncoming vehicle targets and pedestrian targets. This can be achieved in the following ways, without limitation: perform inter-frame difference calculation on consecutive video frames within the corrected blind zone monitoring area to obtain a pixel change amplitude distribution map; perform binarization processing on the pixel change amplitude distribution map to generate a motion region mask; extract connected components from the motion region mask and calculate the geometric feature parameters of each connected component, including the area and aspect ratio of the connected component; track the centroid position change of each connected component in consecutive video frames to calculate the motion vector of each connected component; and determine whether a monitoring target exists based on the geometric feature parameters, motion vectors, preset vehicle shape thresholds, and preset pedestrian shape thresholds.

[0076] S107. If a monitoring target exists, generate a blind zone alarm signal and output it.

[0077] The blind spot alarm signal is an electrical signal generated by the controller after detecting a monitored target to warn the driver of potential danger. This signal can be converted into visual or auditory warnings and is divided into a first-level alarm signal and a second-level alarm signal. The first-level alarm signal is a high-level alarm signal generated by the controller when it detects an oncoming vehicle, used to warn of high-risk blind spot hazards, and is manifested as high-frequency flashing or rapid honking. The second-level alarm signal is a medium-level alarm signal generated by the controller when it detects a pedestrian, used to warn of medium-risk blind spot hazards, and is manifested as low-frequency flashing or voice prompts. High-frequency flashing refers to the alarm indicator flashing at a preset high-frequency frequency, for example, 5 times / second, to quickly attract the driver's attention. Rapid honking refers to the alarm horn sounding at a preset high-frequency rhythm, for example, short continuous honking. Low-frequency flashing refers to the alarm indicator flashing at a preset low-frequency frequency, for example, 2 times / second, to smoothly remind the driver. Voice prompts refer to the playback of preset voice warning content through the vehicle's voice system, for example, "Pedestrian in right blind spot, please be careful."

[0078] Specifically, the controller first identifies the specific type of the monitored target based on the judgment result of step S106, distinguishing between oncoming vehicle targets and pedestrian targets. If it is determined to be an oncoming vehicle target, the controller immediately generates a first-level alarm signal. This signal is a two-channel synchronous output electrical signal. One channel is transmitted to the vehicle's BSD alarm indicator light, driving the indicator light to flash continuously at a preset high frequency. The other channel is transmitted to the vehicle's alarm horn, driving the horn to emit a rapid, continuous sound. The high-frequency flashing and rapid sounding occur simultaneously, forming a strong visual and auditory dual warning until the oncoming vehicle target leaves the blind spot. If it is determined to be a pedestrian target, the controller generates a second-level alarm signal. This signal can be output in a single visual or auditory form, or in a combined form. If it is in visual form, it transmits... The signal is sent to the warning indicator light, which then flashes at a preset low frequency. If the signal is audible, it is transmitted to the vehicle's voice system, which retrieves a pre-stored pedestrian warning voice package and provides a looped voice prompt. The voice prompt includes the location of the blind spot and the type of target, providing accurate alerts. While outputting the alarm signal, the controller continuously receives real-time monitoring footage and tracks the monitored target. If the monitored target leaves the corrected blind spot monitoring area, the controller immediately stops outputting the alarm signal and returns to normal monitoring. If multiple monitored targets enter the blind spot simultaneously, the controller will output alarm signals of the corresponding level. When both oncoming vehicles and pedestrians are present, the first-level alarm signal is output first to ensure that the driver responds to high-risk hazards first.

[0079] Optionally, under normal circumstances, if a monitoring target exists, generating and outputting a blind spot alarm signal can be achieved in the following ways, without limitation: if there is an oncoming vehicle target, a first-level alarm signal is generated and output, which is a high-frequency flashing or rapid honking signal; if there is a pedestrian target, a second-level alarm signal is generated and output, which is a low-frequency flashing or voice prompt signal.

[0080] By adopting the above technical solution, the controller establishes zero-position reference coordinates and delineates the blind spot monitoring area based on a fixed reference object on the vehicle body. It can also detect the positional offset of the detachable camera (including translational offset value and rotational offset angle) in real time. When the positional offset does not exceed the preset loosening threshold, the blind spot monitoring area is corrected through a coordinate compensation matrix. Then, based on the corrected blind spot monitoring area, the monitoring target is detected and an alarm is triggered. This not only realizes the detachable installation of the camera in the BSD alarm device, solving the problem of inconvenient maintenance and replacement of cameras in traditional fixed BSD alarm devices, but also dynamically corrects the monitoring area when the detachable camera has a slight positional offset, avoiding the situation of misalignment, missed detection, or false alarm caused by camera offset, thus ensuring the accuracy and stability of blind spot monitoring. At the same time, by identifying the monitoring target through pixel change features, it can accurately perceive the oncoming vehicles, pedestrians, and other targets in the rear blind spot of the vehicle, and output alarm signals in a timely manner, effectively reducing the collision risk when the vehicle changes lanes or turns.

[0081] The following provides a more detailed description of the process of the method provided in this implementation. Please refer to [link / reference]. Figure 2 This is another flowchart illustrating the alarm method for the detachable BSD alarm device in this application embodiment.

[0082] S201. Monitor the electrical connection status between the fixed base and the detachable camera; when the positioning level signal generated by the detachable camera connecting to the fixed base is detected, output a locking command to the electromagnetic locking mechanism in the fixed base to drive the electromagnetic locking mechanism to mechanically lock the detachable camera; after confirming that the locking is completed, open the wireless video data transmission channel with the detachable camera.

[0083] The electrical connection status refers to the circuit continuity between the fixed base and the detachable camera, indicating the on / off state of the hardware electrical link between them. It includes three states: not connected, intermittent connection, and fully connected. This is the electrical basis for the controller to determine whether the camera is physically in place. The placement level signal is an electrical signal with a preset amplitude generated by the level detection circuit inside the base after the detachable camera and fixed base have completed physical engagement and electrical contact. It is a fixed level between high and low levels (example: high level 3.3V), used to feedback to the controller that the camera has been accurately connected to the base. The locking command is an electrical control signal sent by the controller to the electromagnetic locking mechanism, in the form of a pulse or continuous electrical signal, used to trigger the electromagnetic locking mechanism to perform mechanical locking. Action; The electromagnetic locking mechanism refers to an electromagnetically driven mechanical locking component built into the fixed base. It consists of an electromagnet, a locking tongue, and a latching structure, used to mechanically fix the detachable camera and prevent it from loosening or falling off; Mechanical locking refers to the process in which the electromagnetic locking mechanism, under the action of electromagnetic driving force, engages with the camera's slot through the locking tongue, forming a rigid connection between the detachable camera and the fixed base, used to ensure the mechanical stability of the camera installation; The wireless video data transmission channel refers to a two-way communication link established between the controller and the detachable camera based on wireless communication protocols (examples include WiFi, Bluetooth, and ZigBee), dedicated to transmitting video image data, used to realize the real-time transmission of the camera's captured images to the controller.

[0084] Specifically, the controller first sends a detection command to the electrical detection module of the fixed base to initiate continuous monitoring of the electrical connection status between the fixed base and the detachable camera. This monitoring is achieved by collecting the real-time voltage value of the level detection circuit inside the base. The controller continuously compares the collected voltage value with a preset level judgment threshold. When the operator inserts the detachable camera into the fixed base, and the two complete the physical connection and full contact of the electrical contacts, the level detection circuit inside the base triggers an in-place level signal. After the controller detects the in-place level signal of the preset amplitude, it immediately stops continuous detection and generates a locking command. This command is a pulse electrical signal with a preset duty cycle, which is transmitted to the electromagnetic locking mechanism through the control circuit inside the base. The electromagnet receives a locking command and generates electromagnetic attraction, driving the locking tongue to extend and engage with the matching slot of the detachable camera, thus mechanically locking the camera. Simultaneously, the position detection switch of the electromagnetic locking mechanism triggers a locking completion feedback signal, which is transmitted to the controller in a low-level manner. Upon receiving this feedback signal, the controller confirms that the mechanical locking is complete. Subsequently, the controller sends a channel establishment command to its own wireless communication module and the camera's wireless communication module. Based on pre-stored pairing information and a preset wireless communication protocol, the two complete handshake authentication and establish an encrypted wireless video data transmission channel. After the channel is established, the controller sends a video acquisition start command to the camera, and the camera begins to acquire visual images of the rear of the vehicle.

[0085] S202. Receive the initial monitoring image transmitted by the detachable camera, identify the edge contour line of the fixed reference object of the vehicle body in the initial monitoring image, and record the coordinate position of the edge contour line in the image coordinate system of the initial monitoring image as the zero reference coordinate.

[0086] For details, please refer to step S101, which will not be repeated here.

[0087] S203. Based on the position of the zero-position reference coordinates in the image coordinate system, extend a preset distance towards the rear of the vehicle in the initial monitoring image according to the preset spatial geometric relationship to delineate the blind spot monitoring area.

[0088] For details, please refer to step S102, which will not be repeated here.

[0089] S204. Obtain the current monitoring image transmitted by the detachable camera, extract the real-time edge contour line of the fixed reference object of the vehicle body in the current monitoring image, compare and calculate the coordinate position of the real-time edge contour line with the zero-position reference coordinate, and obtain the position offset of the detachable camera relative to the installation zero position. The position offset includes translation offset value and rotation offset angle.

[0090] For details, please refer to step S103, which will not be repeated here.

[0091] S205. Determine whether the position offset exceeds the preset loosening threshold.

[0092] For details, please refer to step S104, which will not be repeated here.

[0093] S206. If the position offset exceeds the preset loosening threshold, the drive current output to the electromagnetic locking mechanism is increased to the preset high engagement value to increase the mechanical clamping force of the fixed base on the detachable camera. At the same time, an equipment abnormality prompt signal is generated and sent to the vehicle display terminal.

[0094] The preset loosening threshold refers to the critical judgment standard for positional offset of the detachable camera, pre-stored in the controller, used to determine whether abnormal loosening has occurred. It is a combined threshold of translational offset and rotational offset angle, set based on camera installation accuracy, blind spot monitoring error tolerance, and the mechanical structure characteristics of the equipment. For example, an X / Y axis translational offset exceeding 8 pixels or a rotational offset angle exceeding 2° is considered abnormal loosening. The drive current refers to the electrical drive current output by the controller to the electromagnet of the electromagnetic locking mechanism. It is a DC current used to provide energy for magnetic drive of the electromagnet, initially at the normal pull-in current value. The preset high pull-in value refers to the critical drive current value pre-stored in the controller, which is higher than the normal pull-in current of the electromagnetic locking mechanism. This value is based on the hardware performance calibration of the electromagnetic locking mechanism, ensuring that the mechanism does not overheat and suffer damage. To achieve maximum electromagnetic attraction, the example is a standard attraction current of 1A and a preset high attraction value of 1.5A. Mechanical clamping force refers to the mechanical pressing and securing force exerted by the electromagnetic locking mechanism on the detachable camera through structures such as the locking tongue and latches, under the action of the electromagnet's magnetic force. Its magnitude is positively correlated with the drive current. Equipment malfunction warning signal refers to the electrical signal generated by the controller after detecting abnormal looseness of the camera, used to report equipment failure to the driver. It can be converted into text, icons, lights, etc., and includes information such as the location of the loose equipment and the type of failure. Vehicle display terminal refers to the in-vehicle display device used to display vehicle status, equipment failure, and driving information, such as the in-vehicle central control screen, instrument panel display, and BSD system-specific display panel. It is the carrier for conveying equipment malfunction information to the driver.

[0095] Specifically, the controller first retrieves the pre-stored high engagement value of the electromagnetic locking mechanism from its local memory, and simultaneously sends a current adjustment command to its current drive module. This command includes the target current value and the duration of the current output. Upon receiving the command, the current drive module smoothly increases the drive current output to the electromagnet of the electromagnetic locking mechanism from the normal engagement value to the preset high engagement value, and maintains this current value continuously. Under the action of the higher current, the electromagnet generates a stronger electromagnetic attraction, driving the locking tongue, latch, and other mechanical structures to further tighten the engagement part of the detachable camera, increasing the mechanical clamping force of the fixed base on the camera. This attempts to counteract the camera's loosening and displacement by increasing the clamping force, and to reset the camera to a state as close as possible to its installation zero position. While adjusting the drive current, the controller immediately generates a device abnormality warning signal. This signal is a digital signal with a fault code, which includes information about the camera loosening. The controller receives information such as the specific location of the movement (left rear / right rear) and the specific value of the position offset. Then, via the vehicle's CAN bus or a dedicated communication link, the controller transmits the device malfunction warning signal to the vehicle display terminal. Upon receiving the signal, the vehicle display terminal parses the fault code and displays a fault message according to preset display rules, such as displaying the text "Right-side BSD camera is loose, please check immediately" on the central control screen, accompanied by a constant or slow flashing of the instrument panel fault indicator light. Some terminals may also simultaneously play a slight fault warning sound. After completing the above operations, the controller continuously monitors the camera's position offset. If the offset falls back to a preset looseness threshold within a preset time period, the drive current is restored to the normal pull-in value. If the offset still exceeds the threshold, a high pull-in current output is maintained, and the malfunction warning is continuously displayed on the vehicle display terminal until the driver manually inspects and maintains the device.

[0096] S207. If the position offset does not exceed the preset loosening threshold, a coordinate compensation matrix is ​​constructed based on the position offset. The coordinate range of the blind zone monitoring area is corrected by reverse translation and reverse rotation using the coordinate compensation matrix to obtain the corrected blind zone monitoring area.

[0097] For details, please refer to step S105, which will not be repeated here.

[0098] S208. Extract the pixel change features between consecutive video frames within the corrected blind zone monitoring area, and determine whether a monitoring target exists based on the spatial distribution area and motion direction of the pixel change features.

[0099] For details, please refer to step S106, which will not be repeated here.

[0100] S209. If a monitoring target exists, generate a blind zone alarm signal and output it.

[0101] For details, please refer to step S107, which will not be repeated here.

[0102] The controller in the embodiments of this invention is described below from the perspective of hardware processing. Please refer to [link / reference]. Figure 3 This is a schematic diagram of the physical device structure of the controller in an embodiment of this application.

[0103] It should be noted that, Figure 3 The controller structure shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0104] like Figure 3 As shown, the controller includes a CPU 301, which can perform various appropriate actions and processes based on a program stored in the read-only memory ROM 302 or a program loaded from the storage section 308 into the random access memory RAM 303, such as performing the methods described in the above embodiments. The RAM 303 also stores various programs and data required for system operation. The CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An I / O interface 305 is also connected to the bus 304.

[0105] The following components are connected to I / O interface 305: input section 306 including audio input devices, push-button switches, etc.; output section 307 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 308 including a hard disk, etc.; and communication section 309 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 309 performs communication processing via a network such as the Internet. Drive 310 is also connected to I / O interface 305 as needed. Removable media 311, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 310 as needed so that computer programs read from them can be installed into storage section 308 as needed.

[0106] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 309, and / or installed from removable medium 311. When the computer program is executed by CPU 301, it performs the various functions defined in the present invention.

[0107] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0108] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0109] Specifically, the controller in this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, it implements the alarm method of the detachable BSD alarm device provided in the above embodiment.

[0110] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the controller described in the above embodiments; or it may exist independently and not assembled into the controller. The storage medium carries one or more computer programs that, when executed by a processor of the controller, cause the controller to implement the detachable BSD alarm device alarm method provided in the above embodiments.

[0111] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0112] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0113] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. An alarm method for a detachable BSD alarm device, characterized in that, A controller for a detachable BSD alarm device, the detachable BSD alarm device further comprising a fixed base and a detachable camera, the fixed base being disposed on the vehicle body surface, and the detachable camera being connected to the fixed base via a mechanical snap-fit ​​structure, the method comprising: Receive the initial monitoring image transmitted by the detachable camera, identify the edge contour line of the fixed reference object of the vehicle body in the initial monitoring image, and record the coordinate position of the edge contour line in the image coordinate system of the initial monitoring image as the zero reference coordinate. Based on the position of the zero-position reference coordinates in the image coordinate system, a preset distance is extended towards the rear of the vehicle in the initial monitoring image according to a preset spatial geometric relationship to delineate the blind spot monitoring area; The current monitoring image transmitted by the detachable camera is obtained. The real-time edge contour line of the fixed reference object of the vehicle body is extracted from the current monitoring image. The coordinate position of the real-time edge contour line is compared with the zero-position reference coordinate to calculate the position offset of the detachable camera relative to the installation zero position. The position offset includes translation offset value and rotation offset angle. Determine whether the position offset exceeds a preset loosening threshold; If the position offset does not exceed the preset loosening threshold, a coordinate compensation matrix is ​​constructed based on the position offset, and the coordinate range of the blind zone monitoring area is corrected by reverse translation and reverse rotation using the coordinate compensation matrix to obtain the corrected blind zone monitoring area. Extract the pixel change features between consecutive video frames within the corrected blind zone monitoring area, and determine whether a monitoring target exists based on the spatial distribution area and motion direction of the pixel change features; If the monitored target exists, a blind zone alarm signal is generated and output.

2. The method according to claim 1, characterized in that, Before the steps of receiving the initial monitoring image transmitted by the detachable camera, identifying the edge contour line of the fixed reference object of the vehicle body in the initial monitoring image, and recording the coordinate position of the edge contour line in the image coordinate system of the initial monitoring image as the zero-position reference coordinate, the method further includes: Monitor the electrical connection status between the fixed base and the detachable camera; When the positioning level signal generated by the detachable camera being connected to the fixed base is detected, a locking command is output to the electromagnetic locking mechanism in the fixed base to drive the electromagnetic locking mechanism to mechanically lock the detachable camera. After confirming the lock is engaged, the wireless video data transmission channel with the detachable camera is activated.

3. The method according to claim 2, characterized in that, After the step of determining whether the position offset exceeds a preset looseness threshold, the method further includes: If the positional offset exceeds the preset loosening threshold, the drive current output to the electromagnetic locking mechanism is increased to a preset high engagement value to increase the mechanical clamping force of the fixed base on the detachable camera. At the same time, an abnormal device warning signal is generated and sent to the vehicle display terminal.

4. The method according to claim 1, characterized in that, The process involves extracting pixel change features between consecutive video frames within the corrected blind spot monitoring area, and determining the presence of a monitoring target based on the spatial distribution area and motion direction of the pixel change features. The monitoring targets include at least oncoming vehicle targets and pedestrian targets, specifically including: Inter-frame difference calculation is performed on continuous video frames within the corrected blind zone monitoring area to obtain a pixel change amplitude distribution map; The pixel variation amplitude distribution map is binarized to generate a motion region mask; Connected components are extracted from the motion region mask, and geometric feature parameters of each connected component are calculated, including the area and aspect ratio of the connected component. By tracking the centroid position changes of each connected component in consecutive video frames, the motion vector of each connected component is calculated. Based on the geometric feature parameters, the motion vector, the preset vehicle shape threshold, and the preset pedestrian shape threshold, it is determined whether the monitoring target exists.

5. The method according to claim 4, characterized in that, If the monitored target exists, generating and outputting a blind zone alarm signal specifically includes: If the oncoming vehicle is present, a first-level alarm signal is generated and output. The first-level alarm signal is a high-frequency flashing or rapid beeping signal. If the pedestrian target is present, a second-level alarm signal is generated and output. The second-level alarm signal is a low-frequency flashing or voice prompt signal.

6. The method according to claim 1, characterized in that, The step of extending a preset distance towards the rear of the vehicle in the initial monitoring image according to the position of the zero-position reference coordinates in the image coordinate system and a preset spatial geometric relationship to delineate the blind spot monitoring area specifically includes: Using a preset pixel-distance mapping ratio, the preset distance is converted into vertical pixel offset and horizontal pixel offset in the image coordinate system; Using the zero-position reference coordinates as the reference vertex, the coordinates of the far edge of the image in the vertical direction are determined according to the vertical pixel offset, and the coordinates of the outer edge of the image in the horizontal direction are determined according to the horizontal pixel offset. The blind zone monitoring area is formed by the reference vertex, the coordinates of the far boundary, and the coordinates of the outer boundary.

7. The method according to claim 1, characterized in that, The step of constructing a coordinate compensation matrix based on the position offset, and using the coordinate compensation matrix to perform reverse translation and reverse rotation correction on the coordinate range of the blind zone monitoring area to obtain the corrected blind zone monitoring area, specifically includes: A rotation factor is determined based on the rotation offset angle, a translation factor is determined based on the translation offset value, and a coordinate mapping matrix is ​​generated by combining the rotation factor and the translation factor. Extract the preset node coordinates of the blind zone monitoring area; The coordinate transformation calculation is performed on the preset node coordinates using the coordinate mapping matrix to obtain the corrected node coordinates; The boundary of the blind zone monitoring area is reconstructed based on the corrected node coordinates to obtain the corrected blind zone monitoring area.

8. A controller, characterized in that, The controller includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code including computer instructions, and the one or more processors invoking the computer instructions to cause the controller to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the controller, the controller causes the controller to perform the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on the controller, the controller performs the method as described in any one of claims 1-7.

Images