Door control method and device, equipment and storage medium
By using image acquisition and virtual space calculation methods, a three-dimensional point cloud is obtained in the direction of the car door to predict collision risks. This solves the problems of protection lag, detection blind spots and reliability degradation in traditional car door anti-pinch technology, realizes active anti-pinch control, and improves safety and reliability.
Patent Information
- Application Number
- CN202511688333.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-18
- Publication Date
- 2026-01-23
AI Technical Summary
Traditional anti-pinch technology for car doors relies on contact pressure sensors, which suffer from protection lag, detection blind spots, and reliability degradation, and cannot effectively prevent human injury.
Image acquisition technology is used to obtain depth information, which is converted into a 3D point cloud. Combined with virtual space calculation, collision risk is predicted. The spatial relationship between the target and the car door is determined through a virtual model to achieve active anti-pinch control.
It improves the comprehensiveness and accuracy of anti-pinch judgment, avoids human injury caused by delayed protection, and ensures the long-term reliability and safety of the system.
Smart Images
Figure CN121382006A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of vehicle technology, and more particularly to a door control method, device, equipment, and storage medium. Background Technology
[0002] With the continuous development of automotive intelligent technology, the reliability and safety of vehicle door safety systems, as a crucial aspect of human-machine interaction, are receiving increasing attention. During the door closing process, the anti-pinch function is a technology to prevent injuries to passengers' limbs or objects from being pinched. Improving the performance of traditional anti-pinch solutions and developing new intelligent anti-pinch systems have become important research directions in the field of automotive safety.
[0003] Currently, anti-pinch technology for car doors mainly relies on contact pressure sensors installed inside the door frame weatherstripping. This type of technology works by detecting the pressure generated when an object contacts the weatherstripping as the door closes, triggering the anti-pinch action. However, its triggering mechanism has a significant limitation—the sensor can only be activated and perform the anti-pinch operation when a limb or object is actually squeezed and generates sufficiently large pressure.
[0004] This contact-based anti-pinch technology has revealed serious problems in practical applications. The triggering method exhibits significant hysteresis; because it requires a specific pressure threshold to respond, damage may have already occurred to the human body before the sensor is triggered. Summary of the Invention
[0005] This application provides a door control method, device, equipment, and storage medium that can reduce the lag in door control, thereby reducing the impact of the door on the human body.
[0006] To achieve the above objectives, this application adopts the following technical solution: In a first aspect, this application provides a method for controlling a door, including: Acquire a target image with depth information along the direction of the door's movement; The pixels of the target object in the target image are converted into three-dimensional spatial points to obtain the target point cloud; Calculate the shortest spatial distance between the target point cloud and the virtual model of the gate at multiple time points in the virtual space; Based on the shortest spatial distance at multiple moments, calculate the changing trend of the spatial positional relationship between the target point cloud and the virtual model of the gate in the virtual space; Based on the changing trend of the spatial positional relationship between the target point cloud and the virtual model of the door in the virtual space, it is determined whether there is a collision risk between the target point cloud and the virtual model of the door, and the judgment result is obtained. Based on the judgment result, the door is controlled.
[0007] Optionally, the method further comprises: judging whether the shortest spatial distances at the plurality of time instants are all greater than or equal to a first distance threshold value; the calculating of the change trend of the spatial position relationship between the target point cloud and the virtual model of the door in the virtual space according to the shortest spatial distances at the plurality of time instants comprises: in a case where the shortest spatial distances at the plurality of time instants are all greater than or equal to the first distance threshold value, the change trend of the spatial position relationship between the target point cloud and the virtual model of the door in the virtual space is calculated according to the shortest spatial distances at the plurality of time instants.
[0008] Optionally, the method further comprises: in a case where the shortest spatial distances at the plurality of time instants are not all greater than or equal to the first distance threshold value, the collision time is calculated according to the shortest spatial distances at the plurality of time instants; the judging of whether there is a collision risk between the target point cloud and the virtual model of the door according to the collision time and the shortest spatial distances at the plurality of time instants, to obtain a judgment result.
[0009] Optionally, the judging of whether there is a collision risk between the target point cloud and the virtual model of the door according to the collision time and the shortest spatial distances at the plurality of time instants, to obtain a judgment result, comprises: obtaining a number of time instants in which the shortest spatial distances are less than the first distance threshold value in the plurality of time instants; if the collision time is less than a first time threshold value or the number of time instants is greater than or equal to a first number threshold value, a judgment result of existing a collision risk is obtained.
[0010] Optionally, the controlling of the door according to the judgment result comprises: if the judgment result represents that there is a collision risk, the door is controlled to stop moving, the door is controlled to move reversely, or the door is controlled to stop moving first and then controlled to move reversely.
[0011] Optionally, the method further comprises: if the collision time is less than a second time threshold value and the number of time instants is greater than or equal to a second number threshold value, a prompt information is presented; wherein the second time threshold value is greater than the first time threshold value, and the second distance threshold value is greater than the first distance threshold value.
[0012] Optionally, the method further comprises: the door is controlled to keep still.
[0013] In a second aspect, the application provides a door control device, comprising: an acquisition module configured to acquire a target image with depth information in a moving direction of a door; The data processing module is configured to convert pixel points of a target object in the target image into three-dimensional space points to obtain a target point cloud, calculate shortest space distances between the target point cloud and a virtual model of the door at multiple moments in a virtual space, calculate a change trend of a spatial positional relationship between the target point cloud and the virtual model of the door in the virtual space according to the shortest space distances at the multiple moments, and determine whether a collision risk exists between the target point cloud and the virtual model of the door according to the change trend of the spatial positional relationship between the target point cloud and the virtual model of the door in the virtual space to obtain a determination result. The control module is configured to control the door according to the determination result.
[0014] In a third aspect, the present application provides an electronic device including a memory and a processor. The memory stores one or more computer programs including instructions, and when the instructions are executed by the processor, the electronic device performs the method of any one of the first aspect.
[0015] In a fourth aspect, the present application provides a computer-readable storage medium for storing a computer program for executing the method of any one of the first aspect.
[0016] According to the above technical solution, the present application has at least the following beneficial effects: The present application obtains a target image containing depth information in the moving direction of the door, converts target object pixel points into a three-dimensional point cloud, and determines a collision risk in advance through a change trend of a spatial positional relationship between the target point cloud and a virtual model of the door in a virtual space, so that potential dangers can be perceived without physical contact, and the passive response to injury is changed to active prediction and avoidance, thereby avoiding human body damage caused by protective lag from the root.
[0017] Further, the shortest space distances of the target point cloud and the virtual model of the door at multiple moments are calculated, and a collision time is derived to construct a judgment standard with distance and time dual-dimension thresholds. This scheme can accurately adapt to target objects of different volumes and scenes with different door moving speeds, effectively covers small-volume targets that are difficult to detect by traditional technologies, and greatly improves the comprehensiveness and accuracy of anti-pinch judgment.
[0018] Further, when the collision time is less than a first time threshold or the shortest space distance is less than a first distance threshold, the door body is stopped or moved in reverse for emergency protection; when in a second risk interval (the collision time is less than a second time threshold and the shortest space distance is less than a second distance threshold), a prompt information is presented and the door body is controlled to be stationary for early warning. This grading strategy not only ensures rapid protection in emergency scenarios, but also avoids excessive control affecting normal use, balancing safety and convenience.
[0019] The traditional contact technology relies on sensors in the door frame sealing strip, and the aging of the sealing strip can cause the sensitivity of the sensor to decrease, and the protection performance to attenuate after long-term use. The method realizes anti-pinch based on image acquisition and virtual space calculation, the core component does not need to be in mechanical contact with the door, is not affected by factors such as sealing strip aging, environmental temperature and humidity, and can still maintain stable detection and judgment ability during long-term use, thereby ensuring the long-term reliability of the door anti-pinch system.
[0020] It should be understood that the description of technical features, technical solutions, beneficial effects or similar language in this application does not imply that all features and advantages can be achieved in any single embodiment. On the contrary, it can be understood that the description of a feature or beneficial effect means that the specific technical feature, technical solution or beneficial effect is included in at least one embodiment. Therefore, the description of technical features, technical solutions or beneficial effects in this specification does not necessarily refer to the same embodiment. Further, the technical features, technical solutions and beneficial effects described in this embodiment can be combined in any appropriate manner. Those skilled in the art will understand that the embodiments can be implemented without one or more specific technical features, technical solutions or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects can be identified in specific embodiments that do not embody all embodiments. BRIEF DESCRIPTION OF DRAWINGS
[0021] Figure 1 A flowchart of a door control method provided in an embodiment of the present application; Figure 2 A schematic diagram of a door control device provided in an embodiment of the present application; Figure 3 A schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0022] The terms "first", "second", and "third" and the like in the specification and the drawings of the present application are used to distinguish different objects, and are not intended to limit a specific order.
[0023] In the embodiments of the present application, the words "exemplary" or "for example" are used to mean serving as an example, instance, or illustration. Any embodiment or design presented as "exemplary" or "for example" in the embodiments of the present application should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of "exemplary" or "for example" is intended to present concepts in a particular manner. The words "first", "second", and "third" and the like in the specification and the drawings of the present application are used to distinguish different objects, and are not intended to limit a specific order.
[0024] For the sake of clear and concise description of the following embodiments, first, a brief introduction of related art is given: The virtual model of the door is a digital model constructed based on actual structure parameters of the door, and is fully matched with the size and movement track of the real door, can simulate the movement process of the door in the virtual space, and is used for spatial interaction calculation with the target point cloud to judge the collision risk. The embodiments of the application only take the door as an example, and the technical solutions of the application can also be applied to other automatic door scenes.
[0025] The application mainly solves the three problems of protection lag, detection blind area and reliability attenuation existing in the traditional door anti-pinch technology, and the root causes of these problems are all due to the dependence of the traditional technology on the contact type pressure sensor.
[0026] In terms of protection lag, the contact type pressure sensor needs the object to actually contact the door frame sealing strip and generate enough pressure to trigger the anti-pinch action. However, in the process from contact to pressure reaching the trigger threshold, the continuous movement of the door is enough to cause extrusion injury to the human body, resulting in a passive situation that the trigger has already formed injury.
[0027] In terms of detection blind area, due to the arrangement range of the sensor in the sealing strip and the fixed pressure threshold, for small volume and small contact area targets such as children's slender fingers, the generated pressure is difficult to reach the trigger standard, so that the sensor cannot effectively identify such targets, forming a security protection loophole.
[0028] In terms of reliability attenuation, the sensor is integrated with the sealing strip, and the sealing strip is affected by environmental temperature and humidity changes, aging and wear, etc. for a long time, which can cause elastic decline and structural deformation, and then cause the sensor to reduce the adhesion and the perception sensitivity to attenuate, and the protection performance of the anti-pinch system will decrease after long-term use, which cannot stably guarantee the safety of the door use.
[0029] Therefore, the embodiments of the application provide a door control method, which can be executed by a processing device. The processing device can be a terminal or a server. The terminal includes but is not limited to a vehicle terminal, a smart phone, a tablet computer, a notebook computer, a personal digital assistant or a smart wearable device, etc. The server can be a cloud server, for example, a central server in a central cloud computing cluster or an edge server in an edge cloud computing cluster. Of course, the server can also be a server in a local data center. The local data center refers to a data center directly controlled by the user.
[0030] In view of the problems of protection lag, detection blind area and reliability attenuation caused by the dependence of traditional vehicle door anti-pinch technology on contact type pressure sensor, the application breaks out of the traditional logic of triggering protection by physical contact, and constructs an active anti-pinch system through three-dimensional space perception and virtual interaction prediction. First, the image acquisition technology containing depth information is used to convert the target object in the moving direction of the door body into quantifiable three-dimensional point cloud data. Then, a virtual model synchronized with the real door body is built in the virtual space. By calculating the trend of the spatial position relationship between the target point cloud and the virtual door body, the collision risk is judged in advance. Finally, differentiated control is performed according to the risk level, realizing the transformation from passive response to injury to active prediction and avoidance, while avoiding the defect that mechanical sensors are easily affected by the environment, and ensuring long-term use reliability.
[0031] In order to make the technical solutions of the application more clear and easy to understand, the following describes a door control method provided by an embodiment of the application with reference to the accompanying drawings. As shown in the figure, the figure is a flowchart of a door control method provided by an embodiment of the application. The method comprises: Figure 1 S101, the processing device acquires a target image with depth information in the moving direction of the door.
[0032] The moving direction of the door refers to the direction of the movement track of the door during opening or closing, for example, when the door rotates to close around the hinge, the moving direction points to the inside of the vehicle body; when opening, it points to the outside of the vehicle body, which determines the area of image acquisition, ensuring that the space range that the door movement may contact can be covered.
[0033] Depth information refers to the actual distance data between each pixel point in the image and the image acquisition device (such as a camera). Unlike ordinary images that can only present plane color and contour, images containing depth information can reflect the position of the target object in three-dimensional space, for example, it can distinguish between a finger 5 cm away from the camera and a door frame 30 cm away from the camera, which is the data for building a three-dimensional scene.
[0034] The target image refers to the image collected by focusing on the object that may collide with the door body in the moving direction of the door body, rather than indiscriminately capturing the entire scene, but rather covering the space around the movement track of the door, the image needs to contain potential target objects such as human limbs and articles, providing direct visual data for judging collision risk.
[0035] The processing device establishes a data connection with an image acquisition module, which can specifically adopt a special sensing module integrating a high-definition RGB camera and a ToF depth camera. The image acquisition module acquires images in the direction of the movement track of the opening or closing of the vehicle door. The image is not a common plane image, which can clearly present the plane contour and color features of the target object (such as a human hand, personal belongings, etc.), and can also synchronously record the actual spatial distance between each pixel point in the image and the acquisition module, that is, depth information.
[0036] In this process, the design of acquiring images in the direction of the movement of the door body can accurately cover the key spatial area that can be contacted during the movement of the vehicle door, so as to avoid missing potential collision objects due to deviation of the acquisition range. The characteristic of synchronously acquiring depth information effectively breaks the limitation that common plane images can only reflect two-dimensional visual features and cannot reflect spatial distance. In this way, the processing device can acquire complete image information with target visual features and spatial position data, which provides data support for subsequent conversion of the target object into a three-dimensional point cloud and judgment of collision risk in a virtual space.
[0037] In some embodiments, an integrated sensing module is installed on the front fender of one side or both sides of the vehicle. Inside the module, a high-definition RGB camera with a resolution of 1.2 MP or above and a ToF depth camera with VGA resolution (640x480) or HVGA resolution (320x240) are installed in a common optical path or near-axial parallel manner. This design ensures extremely small parallax and high calibration accuracy, provides a hardware foundation for pixel-level image-point fusion, and can accurately capture visual and depth information in the direction of the movement of the vehicle door.
[0038] The module is subjected to strict internal and external parameter joint calibration when it is shipped, so as to obtain an accurate rotation and translation matrix between the RGB camera and the ToF camera, so that the data acquired by the two cameras can be accurately matched. At the same time, the external parameter of the entire module relative to the vehicle coordinate system is also accurately calibrated, so as to ensure the consistency of subsequent three-dimensional space calculation and real vehicle position.
[0039] When the vehicle door is opened and the vehicle speed is zero, the system is activated, and the RGB camera and the ToF camera synchronously capture image frames and depth frames in the direction of the movement of the vehicle door at a frequency of 30 Hz. The image frames record the plane visual features of the target object, and the depth frames record the actual distance of each pixel point from the camera. Subsequently, the ToF raw data is subjected to denoising and filtering processing, for example, by using amplitude-based confidence weighting filtering, bilateral filtering, etc., to remove data noise interference and generate an accurate depth map, thereby providing high-quality data for subsequent three-dimensional space conversion.
[0040] S102, the processing device converts the pixel points of the target object in the target image into three-dimensional space points to obtain a target point cloud.
[0041] A pixel point of a target object refers to the smallest visual unit constituting a target object (such as a human hand or an article) in a target image. Each pixel point contains two-dimensional coordinates and color information and is a basic constituent element of the target object in a planar image.
[0042] A three-dimensional space point refers to a point having clear X, Y, and Z coordinates in a three-dimensional coordinate system after the conversion of a two-dimensional pixel point combined with depth information. The three-dimensional space point can accurately reflect the physical position of a certain position on a target object in a real space.
[0043] A target point cloud is a discrete point cluster formed by a large number of three-dimensional space points of a target object. These points collectively constitute the three-dimensional shape and spatial distribution of the target object and are equivalent to outlining the three-dimensional profile of the target object with an infinite number of space points. The target point cloud is a data form for subsequent analysis of collision risks in a virtual space.
[0044] After obtaining a target image containing depth information, the processing device first identifies and extracts all pixel points corresponding to the target object from the image. These pixel points originally only contain planar coordinates and color information. Subsequently, the processing device uses the internal and external parameters pre-calibrated by the image acquisition module, including the camera focal length, distortion coefficient, and relative position to the vehicle coordinate system, to perform fusion calculation on the two-dimensional coordinates of each pixel point and the depth information (distance from the pixel point to the acquisition module) corresponding to the point, and convert the two-dimensional coordinates into a space point in a three-dimensional coordinate system, that is, to assign the physical coordinates of three dimensions X, Y, and Z to each point. Finally, the data set formed by the collection of these three-dimensional space points, which completely reflects the three-dimensional shape and spatial position of the target object, is the target point cloud.
[0045] The purpose of this conversion process is to convert the abstract visual information in a planar image into quantifiable three-dimensional space data, so that the processing device can accurately restore the real size, shape, and relative position relationship with the door body of the target object in a virtual space, providing a basis for three-dimensional space analysis for subsequent judgment of whether the target object and the door body exist a collision risk.
[0046] In some embodiments, the system runs a lightweight instance segmentation network (such as a lightweight variant of Mask R-CNN). The network can not only identify the bounding box of the target object (such as a human hand) in the image, but also output a pixel-accurate binary mask (Mask) that sets the pixels belonging to the "hand" in the image to 1 and the remaining pixels to 0. Compared with traditional bounding box detection, this pixel-level mask can avoid the interference of background pixels (such as the vehicle body and the ground) within the bounding box on subsequent point cloud extraction and is a key prerequisite for pure target point cloud extraction.
[0047] Then, the system uses the factory-calibrated camera internal and external parameters to back-project each pixel point in the processed ToF depth map into a three-dimensional space point, forming a complete scene point cloud; and uses the binary mask obtained by the above instance segmentation as a filter to only retain the three-dimensional points in the image whose corresponding positions are 1 (i.e., belonging to the target hand), and finally obtains a real-time, high-density three-dimensional point cloud model containing only the target hand and no background interference.
[0048] In S103, the processing device calculates the shortest spatial distance between the target point cloud and the virtual model of the door at multiple moments in the virtual space, and calculates the change trend of the spatial position relationship between the target point cloud and the virtual model of the door in the virtual space according to the shortest spatial distance at the multiple moments.
[0049] The virtual space refers to a digital simulation environment constructed by the processing device, which can simulate the coordinate system and object motion law of the real physical space, and provides a digital stage for interactive calculation of the target point cloud and the virtual model of the door, without relying on the real physical scene to complete the risk analysis.
[0050] The virtual model of the door is a digital model constructed based on the structural parameters (such as size, shape, and hinge position) and motion characteristics (such as opening / closing trajectory and speed) of the real door, and its motion state in the virtual space is completely synchronized with the real door, which can accurately simulate the actual movement process of the door.
[0051] In some embodiments, the system preloads high-precision three-dimensional CAD models of the vehicle body (especially the B-pillar and door frame part) and the rear door in the system memory, which are denoted as static model M_body and dynamic model M_door, respectively. At the same time, the system reads the value θ(t) of the rear door angle sensor in real time through the CAN bus, and the processing unit updates the pose (position and attitude) of the dynamic model M_door in the vehicle coordinate system in real time according to θ(t), forms a virtual dynamic door model M_door(t) that is completely synchronized with the motion of the real door, and constructs a virtual simulation environment consistent with the physical world.
[0052] The collision risk judgment refers to an analysis process of judging whether the target point cloud and the virtual model of the door will physically contact at a future moment by calculating the change trend of the spatial position relationship between the two in the virtual space, and is the basis for active anti-pinch.
[0053] The minimum spatial distance at multiple time points refers to the minimum distance between all three-dimensional points in the target point cloud and the surface of the virtual door model calculated by the processing device at fixed time intervals (e.g., every 10 milliseconds) within the continuous movement cycle of the door opening or closing. This forms a set of distance data that changes over time and can be used to calculate the trend of the spatial position relationship between the two. For example, within the continuous movement cycle of the door opening or closing, the processing device calculates the minimum spatial distance between all points in the target point cloud and the virtual dynamic door model (especially the closed edge of the door) at fixed time intervals, such as every 10 milliseconds. The interval needs to be small enough to cover the entire movement and capture the continuous position change. A high-efficiency near-field collision detection algorithm is used, which does not require the construction of complex geometric bodies and can directly iterate on the point set and convex hull. The algorithm is fast and accurate, and can calculate the minimum spatial distance in real time. After multiple calculations, a set of data containing time points and corresponding minimum spatial distances is formed, which can be used to calculate the trend of the spatial position relationship between the two.
[0054] The processing device first imports the target point cloud and the virtual door model into a unified virtual space, ensuring that both follow the same three-dimensional coordinate system and movement rules.
[0055] Within the continuous movement cycle of the door opening or closing, the processing device performs multiple distance calculations at pre-set fixed time intervals. The interval needs to be small enough to cover the entire movement and capture the continuous position change. Each time the calculation is performed, all three-dimensional points in the target point cloud are traversed, and the straight-line distance from each point to the surface of the virtual door model is calculated. The minimum value from these distances is then selected, which is the minimum spatial distance at that time.
[0056] By repeating this operation, the processing device obtains a set of data containing time points and corresponding minimum spatial distances, such as 10 cm at time t1, 8 cm at time t2, and 6 cm at time t3. This forms a data set that can be used to calculate the dynamic change in the spatial position relationship between the two.
[0057] This set of time points and corresponding minimum spatial distance data is like a dynamic recording curve that clearly shows the change in the spatial relationship between the target point cloud and the virtual door model during the movement process. For example, from time t1 to time t3, the minimum spatial distance gradually shortens from 10 cm to 6 cm, directly reflecting that the two are continuously approaching each other. If the minimum spatial distance continues to shorten in subsequent data, it indicates that the approaching trend is still continuing.
[0058] If the minimum spatial distance increases, it indicates that the two are starting to move away.
[0059] In some examples, the shortest spatial distances at 5 continuous time points are 3.2 cm, 2.7 cm, 2.1 cm, 1.4 cm and 0.8 cm respectively, the time interval is 0.1 s, the distance change amount at adjacent time points can be calculated first, and then the distance change rate can be calculated. For example, the calculation result of the distance change rate can be +5 cm / s, +6 cm / s, +7 cm / s and +6 cm / s, wherein the negative sign indicates that the distance increases, and the positive sign indicates that the distance decreases. The shortest spatial distance between the target point cloud and the virtual model of the door continues to decrease, and the distance change rate is stable at 5-7 cm / s, indicating that the spatial positional relationship between the two is in a rapid approaching trend.
[0060] In some examples, the processing device can first determine whether the shortest spatial distances at the plurality of time points are all greater than or equal to the first distance threshold; in the case that the shortest spatial distances at the plurality of time points are all greater than or equal to the first distance threshold, the change trend of the spatial positional relationship between the target point cloud and the virtual model of the door in the virtual space is calculated according to the shortest spatial distances at the plurality of time points.
[0061] The change trend of the spatial positional relationship is calculated only for the case that the shortest spatial distances at the plurality of time points are all not less than the first distance threshold, that is, for the case that there is no need for emergency response, which can not only avoid unnecessary calculation consumption and improve the operation efficiency of the processing device, but also make the trend calculation more targeted and ensure that the subsequent collision risk judgment and the control of the door are more in line with the actual scene requirements. In this scene, there is no collision risk in the case of a long distance, but the sudden situation needs to be considered, and the application reflects whether there is a sudden situation (for example, the hand of the user suddenly approaches the door body, or the door body suddenly approaches the hand of the user) through the change trend of the spatial positional relationship.
[0062] S104, the processing device determines whether there is a collision risk between the target point cloud and the virtual model of the door according to the change trend of the spatial positional relationship between the target point cloud and the virtual model of the door in the virtual space, and obtains a judgment result.
[0063] As introduced before, the change trend can be represented by the distance change rate. In the embodiment of the application, the processing device can acquire a first continuous number of continuous distance change rates exceeding a first change rate threshold; if the first continuous number is greater than a continuous number threshold, it is determined that there is a collision risk.
[0064] For example, the first change rate threshold can be +3 cm / s, and the continuous number threshold can be 3. In the above example, the calculation result of the distance change rate can be +5 cm / s, +6 cm / s, +7 cm / s and +6 cm / s, that is, the first continuous number is 4, which exceeds the continuous number threshold (3), that is, it is considered that there is a collision risk between the target point cloud and the virtual model of the door, and a judgment result that there is a collision risk is obtained.
[0065] In some embodiments, in a case where the minimum spatial distance at multiple time instants is greater than or equal to a first distance threshold, a collision time is calculated according to the minimum spatial distance at the multiple time instants; whether there is a collision risk between the target point cloud and the virtual model of the door is judged according to the collision time and the minimum spatial distance at the multiple time instants, and a judgment result is obtained.
[0066] Further, in order to improve the accuracy of the judgment, the judgment can also be combined with the index of the collision time. The same is that the minimum spatial distance is shortened from 10 cm to 6 cm in 300 ms, and the speed of approaching is completely different from that in 200 ms, and the urgency of future collision is also completely different.
[0067] Therefore, based on the set of dynamic minimum spatial distance data, a key index of energyized collision urgency can be further calculated to provide a more accurate basis for subsequent risk judgment.
[0068] Specifically, the processing device calculates a collision time according to the minimum spatial distance at multiple time instants; whether there is a collision risk between the target point cloud and the virtual model of the door is judged according to the collision time and the minimum spatial distance at the multiple time instants, and a judgment result is obtained.
[0069] The collision time is an index of quantifying the collision urgency, which is derived based on the minimum spatial distance between the target point cloud and the virtual model of the door at multiple time instants, and the relative motion speed of the two. If the current motion state is maintained, the estimated time required for the two to physically contact is the estimated time required for the two to physically contact. The calculation expression of the collision time is:
[0070] Wherein, represents the collision time, that is, the estimated collision remaining time at time, represents the minimum spatial distance, represents the relative approaching speed.
[0071] The calculation expression of the collision time is:
[0072] Wherein, represents the distance change rate of the minimum spatial distance, which is the derivative of with respect to time, reflecting the change trend of the minimum spatial distance with respect to time. If , it means that the minimum spatial distance is shortened (the collision risk is increased); if , it means that the minimum spatial distance is increased (the collision risk is reduced).
[0073] For example, assume that a car door is closing, and the shortest spatial distance between a target point cloud (such as a hand) and a virtual model of the door changes over time as follows: At t0=0 seconds, the shortest spatial distance is 20 cm; at t1=0.1 seconds, the shortest spatial distance is 12 cm; .
[0074] From t0to t1, the shortest spatial distance changes by 8 10= cm, and the time changes by 0.1 0=0.1 seconds, so:
[0075] Take t2as an example: at this time, , the shortest spatial distance is 10 cm / s, so: seconds This means that if the current closing speed of 20 cm / s is maintained, the hand will collide with the car door after 0.3 seconds, i.e., at t2+0.3=0.2+0.3=0.5 seconds.
[0076] The processing device can obtain the number of times when the shortest spatial distance is less than a first distance threshold in multiple time points, and if the collision time is less than a first time threshold or the number of times is greater than or equal to a first number threshold, a determination result that there is a collision risk is obtained.
[0077] The first time threshold is a critical value of the collision time, used to define the time boundary of the high-risk collision risk. For example, the first time threshold can be set to 100 milliseconds, and when the collision time is less than 100 milliseconds, it means that the collision is extremely urgent.
[0078] The first distance threshold is a critical value of the shortest spatial distance, used to define the spatial boundary of the high-risk collision risk. For example, the first distance threshold can be set to 1.5 cm, and when the shortest spatial distance at a certain time is less than 1.5 cm, it means that the target and the door body are extremely close.
[0079] In the collision risk judgment logic, the processing device sets double triggering conditions, and as long as any one of the conditions is met, it is determined that there is a collision risk: Condition one: the collision time is less than the first time threshold. For example, the first time threshold is set to 100 milliseconds, and if the calculated collision time is 80 milliseconds, it means that the target and the door body will collide in a very short time, and are in a high-risk state.
[0080] Condition two: the number of times is greater than or equal to a first number threshold. For example, the first distance threshold is set to 1.5 cm, and the first number threshold is 4. If, in a certain judgment, there are 2 shortest space distances of 1 cm and 3 shortest space distances of 1.2 cm, it indicates that the target has approached the door body multiple times and physical contact is about to occur.
[0081] The processing device ensures timely identification of risks in the high-risk stage before the collision occurs by monitoring the two dimensions of time urgency (collision time) and spatial proximity (shortest space distance) at the same time, and provides a basis for subsequent door control actions.
[0082] If the collision time is less than a second time threshold and the number of times is greater than or equal to a second number threshold, a prompt information is presented; wherein the second time threshold is greater than the first time threshold, and the second number threshold is less than the first number threshold.
[0083] The second time threshold is a secondary critical value of the collision time, used to define the time boundary of the non-urgent but vigilant collision risk, and the second time threshold is greater than the first time threshold, for example, the first time threshold is 100 ms, and the second time threshold can be set to 300 ms, representing that the collision is not extremely urgent, but needs to be warned in advance.
[0084] The second number threshold is a secondary critical value of the number of shortest space distances, used to define the spatial boundary of the non-urgent but vigilant collision risk, and the second number threshold is less than the first distance threshold, for example, the first number threshold is 4, and the second number threshold can be set to 2, representing that the target has approached the door body twice but has not reached the high-risk contact range.
[0085] The prompt information is used to deliver a warning signal to the user that there is a potential collision risk near the door, and the form can include visual prompts such as flashing of the door perimeter indicator light, auditory prompts such as warning sounds from the buzzer, and tactile prompts such as vibration of the in-vehicle control panel, etc., and does not directly control the door movement, but only serves as a risk prompt.
[0086] When the two conditions are met at the same time, the processing device will trigger the prompt information: Condition one: the collision time is less than the second time threshold, for example, the second time threshold is set to 300 ms, and if the calculated collision time is 200 ms, it indicates that the target and the door body have a collision possibility, but still have a certain reaction time, and have not reached the degree of needing to control the door body urgently.
[0087] Condition two: the number of times is greater than or equal to the second number threshold, for example, the second number threshold is set to 2, and if, in a certain judgment, there are 2 shortest space distances of 1 cm, it indicates that the target has entered the approach range that needs to be vigilant, but has not reached the high-risk state of extremely close.
[0088] S105, the processing device controls the door according to the judgment result.
[0089] If the judgment result represents that there is a collision risk, the door is controlled to stop moving, the door is controlled to move reversely, or the door is controlled to stop moving first and then move reversely.
[0090] Controlling the door to stop moving means that the processing device sends an instruction to the driving system of the door to immediately cut off the power output, so that the door stops the current opening or closing action, and avoids the door continuing to move to squeeze the target.
[0091] Controlling the door to move reversely means that on the basis of the door stopping, the processing device instructs the driving system to change the movement direction of the door, for example, the door is originally closing, and at this time, the door is controlled to open to quickly pull away from the target and completely eliminate the collision risk.
[0092] When the processing device determines that there is a collision risk, in order to avoid collision damage in an emergency, one of the following three types of door control actions is executed according to the emergency degree of the risk and the system design: If the collision risk is extremely urgent (such as the collision time is extremely short), the door can be directly controlled to move reversely to quickly pull away; if it is necessary to stabilize the state of the door before performing the reverse action, the door can be controlled to stop moving first and then controlled to move reversely; in some scenarios, the door can also be controlled to stop moving only to prevent the door from continuing to move, and the risk is handled subsequently in combination with manual intervention or other logic.
[0093] The purpose of these control methods is to change the movement state of the door in time to avoid physical contact between the target (such as a hand or an object) and the door, and to achieve the final protection effect of active anti-pinch.
[0094] The method of controlling the door also includes controlling the door to remain stationary.
[0095] When the collision time is greater than the first time threshold but less than the second time threshold, and the shortest spatial distance is greater than the first distance threshold but less than the second distance threshold, the processing device will first trigger a warning prompt. If the user fails to respond to the warning in time, the processing device will temporarily lock the state of the door by controlling the door to remain stationary, neither continuing to perform the original movement to avoid approaching the target nor moving reversely, thereby reserving sufficient time for the user to adjust the position of the target, while continuously cooperating with the prompt information to strengthen the risk notification and ensure that the user clearly perceives the current safety requirement.
[0096] When the processing device stops moving by controlling the door to avoid the immediate collision risk, the processing device further monitors the position of the target object through the image acquisition module. If it is found that the target object is still near the door and has not completely left the risk area, if the opening or closing movement of the door is immediately resumed, the collision risk is likely to be triggered again. Therefore, the processing device further controls the door to remain stationary, continuously locks the door state, until the image acquisition module confirms that the target object has completely left the risk area, or an explicit operation instruction (such as pressing the door opening key or the door closing key again) is received from the user, and then the stationary locking of the door is released, thereby fundamentally avoiding the occurrence of secondary collision risk.
[0097] Based on the above content description, the present application has the following beneficial effects: The present application obtains a target image containing depth information in the moving direction of the door, converts the target object pixel points into three-dimensional point clouds, and judges the collision risk in a virtual space in advance, so that potential dangers can be perceived without physical contact, and the passive response to injury is changed to active prediction and avoidance, thereby fundamentally avoiding human body damage caused by protection lag.
[0098] Further, the shortest spatial distance between the target point cloud at multiple moments and the virtual model of the door is calculated, and the collision time is derived, so as to construct a judgment standard with distance and time two-dimensional threshold. This logic can accurately adapt to target objects of different volumes and scenes with different door moving speeds, effectively cover small-volume targets that cannot be detected by traditional technology, and greatly improve the comprehensiveness and accuracy of anti-pinch judgment.
[0099] Further, when the collision time is less than the first time threshold or the shortest spatial distance is less than the first distance threshold, the door body is stopped or moved in the opposite direction for emergency protection; when in the second risk interval (the collision time is less than the second time threshold and the shortest spatial distance is less than the second distance threshold), a prompt information is presented and the door body is controlled to be stationary for early warning. This grading strategy not only ensures rapid protection in emergency scenarios, but also avoids excessive control affecting normal use, balancing safety and convenience.
[0100] The traditional contact type technology relies on sensors in the door frame sealing strip, and the aging of the sealing strip can cause the sensitivity of the sensor to decrease, and the protection performance to decay after long-term use. The present method realizes anti-pinch based on image acquisition and virtual space calculation, and the core components do not need to be in mechanical contact with the door, and are not affected by factors such as aging of the sealing strip, environmental temperature and humidity, and can still maintain stable detection and judgment ability during long-term use, thereby ensuring the long-term reliability of the door anti-pinch system.
[0101] The above is combined Figure 1 The door control method provided by the embodiments of the present application is described in detail, and the device and the equipment provided by the embodiments of the present application will be introduced below with reference to the accompanying drawings.
[0102] AsFigure 2 As shown in the figure, the figure is a schematic diagram of a door control device provided by an embodiment of the application, and the device comprises: The acquisition module 201 is configured to acquire a target image with depth information in a moving direction of a door. The data processing module 202 is configured to convert a pixel point of a target object in the target image into a three-dimensional space point to obtain a target point cloud, calculate a shortest space distance between the target point cloud and a virtual model of the door at multiple moments in a virtual space, calculate a change trend of a spatial position relationship between the target point cloud and the virtual model of the door in the virtual space according to the shortest space distances at the multiple moments, and determine whether there is a collision risk between the target point cloud and the virtual model of the door according to the change trend of the spatial position relationship between the target point cloud and the virtual model of the door in the virtual space to obtain a determination result. The control module 203 is configured to control the door according to the determination result.
[0103] Optionally, the data processing module 202 is further configured to determine whether the shortest space distances at the multiple moments are all greater than or equal to a first distance threshold, and specifically configured to calculate the change trend of the spatial position relationship between the target point cloud and the virtual model of the door in the virtual space according to the shortest space distances at the multiple moments in a case where the shortest space distances at the multiple moments are all greater than or equal to the first distance threshold.
[0104] Optionally, the data processing module 202 is further configured to calculate a collision time according to the shortest space distances at the multiple moments in a case where the shortest space distances at the multiple moments are not all greater than or equal to the first distance threshold, and determine whether there is a collision risk between the target point cloud and the virtual model of the door according to the collision time and the shortest space distances at the multiple moments to obtain the determination result.
[0105] Optionally, the data processing module 202 is specifically configured to acquire a number of moments at which the shortest space distances are less than the first distance threshold among the multiple moments, and obtain the determination result of the collision risk if the collision time is less than a first time threshold or the number of moments is greater than or equal to a first number threshold.
[0106] Optionally, the control module 203 is specifically configured to control the door to stop moving, control the door to move reversely, or control the door to stop moving first and then control the door to move reversely if the determination result represents the collision risk.
[0107] Optionally, the data processing module 202 is further configured to present prompt information if the collision time is less than a second time threshold and the number of moments is greater than or equal to a second number threshold, wherein the second time threshold is greater than the first time threshold, and the second distance threshold is greater than the first distance threshold.
[0108] Optionally, the control module 203 is further configured to control the door to remain stationary.
[0109] The control device of the door according to the embodiments of the present application can correspond to the method described in the embodiments of the present application, and the above-mentioned other operations and / or functions of each module / unit of the control device of the door are respectively implemented to realize the corresponding flow of each method in the embodiments of the present application. For brevity, the details are not described herein. Figure 1 The corresponding flow of each method in the embodiments of the present application, for brevity, is not described herein.
[0110] The embodiments of the present application also provide an electronic device. As shown in the Figure 3 The figure is a schematic diagram of an electronic device provided by the embodiments of the present application. The electronic device 300 includes a bus 301, a processor 302, a communication interface 303 and a memory 304. The processor 302, the memory 304 and the communication interface 303 communicate through the bus 301.
[0111] The bus 301 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience, only one thick line is used to represent the bus in the figure, but it does not mean that there is only one bus or only one type of bus. Figure 3
[0112] The processor 302 can be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP) processor, etc.
[0113] The communication interface 303 is configured to communicate with the outside.
[0114] The memory 304 can include a volatile memory (volatile memory), such as a random access memory (RAM). The memory 304 can also include a non-volatile memory (non-volatile memory), such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD) or a solid state drive (SSD).
[0115] The executable code is stored in the memory 304, and the processor 302 executes the executable code to perform the control method of the door.
[0116] Specifically, in the case of the embodiment shown in the figure, and Figure 2 In the case of the embodiment shown in the figure, and Figure 2 In the case of the embodiment shown in the figure, and Figure 2 The software or program code required for the functions of each module / unit in the foregoing embodiment can be stored in the memory 304 in part or in whole. The processor 302 executes the program code corresponding to each unit stored in the memory 304 to perform the control method of the door.
[0117] The embodiment of the present application also provides a computer readable storage medium. The computer readable storage medium can be any available medium or data center containing one or more available media that can be stored by an electronic device. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk), etc. The computer readable storage medium includes instructions for instructing an electronic device to perform the control method of the door.
[0118] The embodiment of the present application also provides a computer program product, which includes one or more computer instructions. When the computer instructions are loaded and executed on an electronic device, all or part of the processes or functions described in the embodiment of the present application are generated.
[0119] The computer instructions can be stored in a computer readable storage medium, or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website, computer or data center to another website, computer or data center through wired (such as coaxial cable, optical fiber, digital subscriber line) or wireless (such as infrared, wireless, microwave, etc.) mode.
[0120] The computer program product is executed by a computer, and the computer executes any of the control methods of the door. The computer program product can be a software installation package, and when any of the control methods of the door is needed, the computer program product can be downloaded and executed on the computer.
[0121] The description of the flow or structure corresponding to each of the above figures has its own emphasis, and the parts not described in detail in a certain flow or structure can be referred to the related description of other flows or structures.
[0122] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this, any change or replacement within the technical scope disclosed by the present application should be covered in the protection scope of the present application.
Claims
1. A method for controlling a door, characterized in that, The method includes: Acquire a target image with depth information along the direction of the door's movement; The pixels of the target object in the target image are converted into three-dimensional spatial points to obtain the target point cloud; Calculate the shortest spatial distance between the target point cloud and the virtual model of the gate at multiple time points in the virtual space; Based on the shortest spatial distance at multiple moments, calculate the changing trend of the spatial positional relationship between the target point cloud and the virtual model of the gate in the virtual space; Based on the changing trend of the spatial positional relationship between the target point cloud and the virtual model of the door in the virtual space, it is determined whether there is a collision risk between the target point cloud and the virtual model of the door, and the judgment result is obtained. Based on the judgment result, the door is controlled.
2. The method according to claim 1, characterized in that, The method further includes: Determine whether the shortest spatial distance at multiple time points is greater than or equal to a first distance threshold; The step of calculating the changing trend of the spatial positional relationship between the target point cloud and the virtual model of the gate in the virtual space based on the shortest spatial distance at multiple times includes: When the shortest spatial distance at multiple times is greater than or equal to the first distance threshold, the changing trend of the spatial positional relationship between the target point cloud and the virtual model of the gate in the virtual space is calculated based on the shortest spatial distance at multiple times.
3. The method according to claim 2, characterized in that, The method further includes: If the shortest spatial distance at multiple times is not uniformly greater than or equal to a first distance threshold, the collision time is calculated based on the shortest spatial distance at the multiple times. Based on the collision time and the shortest spatial distance at multiple moments, it is determined whether there is a collision risk between the target point cloud and the virtual model of the door, and the determination result is obtained.
4. The method according to claim 3, characterized in that, The step of determining whether there is a collision risk between the target point cloud and the virtual model of the door based on the collision time and the shortest spatial distance at multiple times, and obtaining the determination result, includes: Obtain the number of times when the shortest spatial distance is less than a first distance threshold among multiple time points; If the collision time is less than the first time threshold or the number of moments is greater than or equal to the first number threshold, then a judgment result indicating that there is a collision risk is obtained.
5. The method according to claim 4, characterized in that, The step of controlling the door based on the judgment result includes: If the judgment result indicates that there is a collision risk, control the door to stop moving, control the door to move in the opposite direction, or first control the door to stop moving and then control the door to move in the opposite direction.
6. The method according to claim 4, characterized in that, The method further includes: If the collision time is less than the second time threshold and the number of moments is greater than or equal to the second number threshold, a prompt message will be displayed. Wherein, the second time threshold is greater than the first time threshold, and the second distance threshold is greater than the first distance threshold.
7. The method according to claim 6, characterized in that, The method further includes: Control the door to remain stationary.
8. A door control device, characterized in that, The device includes: The acquisition module is used to acquire a target image with depth information along the direction of the door's movement; The data processing module is used to convert the pixels of the target object in the target image into three-dimensional spatial points to obtain a target point cloud; calculate the shortest spatial distance between the target point cloud and the virtual model of the door at multiple moments in virtual space; calculate the changing trend of the spatial positional relationship between the target point cloud and the virtual model of the door in virtual space based on the shortest spatial distance at multiple moments; and determine whether there is a collision risk between the target point cloud and the virtual model of the door based on the changing trend of the spatial positional relationship between the target point cloud and the virtual model of the door in virtual space, and obtain a judgment result. The control module is used to control the door based on the judgment result.
9. An electronic device, characterized in that, Including memory and processor; The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the electronic device performs the method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program for performing the method as described in any one of claims 1 to 7.