Methods, apparatus, equipment, and media for detecting vehicle cut-in based on monocular images

Abstract

This disclosure relates to a method, apparatus, device, and medium for detecting a vehicle cutting into another vehicle based on monocular images. The method includes: acquiring multiple consecutive frames of target detection images, each including a detection frame of a vehicle in an adjacent lane; acquiring a first ranging point on the detection frame of the vehicle in each frame; for each frame, obtaining relative motion parameters between the vehicle and the driver based on the first ranging point; and, if the vehicle is determined to be approaching based on the relative lateral distance and relative lateral velocity corresponding to the multiple consecutive frames, calculating the intersection of three planes for every two consecutive frames to obtain an updated first ranging point, and determining whether the vehicle is cutting into another vehicle based on the updated first ranging point. According to embodiments of this disclosure, the accuracy of ranging detection can be improved, ranging detection errors reduced, and vehicle driving safety enhanced.

Description

Technical Field

[0001] This disclosure relates to the field of image recognition technology, and in particular to a method, apparatus, device and medium for detecting vehicle cut-in based on monocular images. Background Technology

[0002] While driving, it is necessary to be constantly aware of whether vehicles on both sides will enter the current driving lane, thereby affecting driving safety. Therefore, it is necessary to measure the distance to vehicles on both sides during the driving process.

[0003] In related technologies, when measuring distances between vehicles on both sides, the center of the bottom edge of the detection frame of the vehicles on both sides is used as the distance measuring point. However, using this distance measuring point for detection is inaccurate, which may lead to misjudgment that vehicles on both sides have entered the current driving lane, thus affecting driving safety. Summary of the Invention

[0004] To address the aforementioned technical problems, this disclosure provides a method, apparatus, device, and medium for detecting vehicle intrusion based on monocular images.

[0005] In a first aspect, this disclosure provides a method for detecting a vehicle cutting in front based on monocular images, including:

[0006] Acquire consecutive multi-frame target detection images, each including the detection bounding box of the vehicle ahead in the adjacent lane;

[0007] Obtain the first ranging point on the detection box of the front vehicle in each frame of the target detection image. The first ranging point is the center point of the bottom edge of the rear end of the front vehicle.

[0008] For each frame of target detection image, the relative motion parameters between the vehicle in front and the vehicle itself are obtained based on the first ranging point. The relative motion parameters include the relative lateral distance and relative lateral velocity of the vehicle in front.

[0009] When it is determined that the vehicle in front has a tendency to approach based on the relative lateral distance and relative lateral velocity corresponding to multiple consecutive target detection images, the first distance measurement point is updated by calculating the intersection of three planes for every two consecutive target detection images in the multiple target detection images, and the vehicle in front has a tendency to cut in is determined based on the updated first distance measurement point.

[0010] Secondly, this disclosure provides another method for detecting vehicle cut-in based on monocular images, including:

[0011] The first acquisition module is used to acquire multiple consecutive target detection images, the target detection images including the detection box of the vehicle in front located in the adjacent lane;

[0012] The second acquisition module is used to acquire the first ranging point on the detection box of the front vehicle in each frame of the target detection image. The first ranging point is the center point of the bottom edge of the rear end of the front vehicle.

[0013] The first processing module is used to obtain the relative motion parameters between the vehicle in front and the vehicle itself based on the first ranging point for each frame of target detection image. The relative motion parameters include the relative lateral distance and relative lateral velocity of the vehicle in front.

[0014] The second processing module is used to calculate the intersection of three planes in every two consecutive target detection images in the multi-frame target detection images to obtain the updated first ranging point when it is determined that the vehicle in front has a tendency to approach based on the relative lateral distance and relative lateral velocity corresponding to the target detection images in consecutive frames. Based on the updated first ranging point, it is used to determine whether the vehicle in front has a tendency to cut in.

[0015] Thirdly, this disclosure provides a vehicle cutting-in detection device based on monocular images, comprising:

[0016] Fourthly, this disclosure provides another vehicle cutting-in detection device based on monocular images, including:

[0017] Fifthly, this disclosure provides a vehicle cutting-in detection device based on monocular images, comprising:

[0018] processor;

[0019] Memory, used to store executable instructions;

[0020] The processor is used to read executable instructions from memory and execute the executable instructions to implement the first aspect or the second aspect of the monocular image-based vehicle intrusion detection method.

[0021] In a sixth aspect, this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to implement the monocular image-based vehicle intrusion detection method of the first or second aspect.

[0022] The technical solution provided in this disclosure has the following advantages compared with the prior art:

[0023] The method, apparatus, device, and medium for detecting a vehicle cutting into another vehicle based on monocular images disclosed herein can acquire a series of multi-frame target detection images including the detection frame of a vehicle in an adjacent lane. Then, a first ranging point on the detection frame of the vehicle in each frame of the target detection image is acquired. This first ranging point is the center point of the rear bottom edge of the vehicle. For each frame of the target detection image, relative motion parameters, including relative lateral distance and relative lateral velocity, are obtained between the vehicle and the other vehicle based on the first ranging point. Finally, if it is determined that the vehicle is approaching based on the relative lateral distance and relative lateral velocity corresponding to the series of multi-frame target detection images, the relative motion parameters between the vehicle and the other vehicle are obtained for each consecutive two frames of the multi-frame target detection images. The frame target detection image is used to calculate the intersection of three planes to obtain the updated first ranging point. Based on the updated first ranging point, it is determined whether the vehicle in front has a tendency to cut in. Thus, the relative motion parameters, namely the relative lateral distance and the relative lateral velocity, can be obtained first through the first ranging point. Then, if the vehicle in front is determined to have a tendency to approach based on the relative lateral distance and the relative lateral velocity, the updated first ranging point is obtained. Finally, the updated first ranging point is used to determine whether the vehicle in front has a tendency to cut in. Therefore, the distance detection of the vehicle in front can be performed by using the first ranging point and the updated first ranging point, thereby improving the accuracy of distance detection, reducing distance detection errors, and improving vehicle driving safety. Attached Figure Description

[0024] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0025] Figure 1 A schematic flowchart illustrating a method for detecting vehicle cut-in based on monocular images provided in this disclosure embodiment;

[0026] Figure 2 This is a schematic diagram of the structure of a target detection image provided in an embodiment of the present disclosure;

[0027] Figure 3 A schematic diagram of another vehicle cutting-in detection device based on monocular images provided in this disclosure embodiment;

[0028] Figure 4 This is a schematic diagram of the structure of a vehicle cutting in front based on monocular images, provided in an embodiment of this disclosure. Detailed Implementation

[0029] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0030] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0031] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0032] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0033] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0034] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0035] While driving, it is necessary to be constantly aware of whether vehicles on both sides will enter the current driving lane, thereby affecting driving safety. Therefore, it is necessary to measure the distance to vehicles on both sides during the driving process.

[0036] In related technologies, when detecting distances between vehicles on both sides, the center of the bottom edge of the detection frame for each vehicle is used as the distance measurement point. When the vehicle is far away or directly in front, the detection frame only includes the rear of the vehicle, not the sides, and the center of the bottom edge of the detection frame is the center of the rear of the vehicle, so there is no lateral deviation in the distance measurement result. However, when the vehicle is on either side, the detection frame includes not only the rear of the vehicle but also the sides. If the center of the bottom edge of the detection frame is used as the distance measurement point, the distance measurement point will be closer to the vehicle than the actual center of the rear of the vehicle due to the influence of the vehicle's sides in the detection frame. This creates the illusion of approaching and cutting in, which may lead to misjudgment that vehicles on both sides have entered the current driving lane, affecting driving safety.

[0037] To address the aforementioned problems, this disclosure provides a method, apparatus, device, and medium for detecting forward vehicle cut-in based on monocular images. The following is in conjunction with… Figure 1 The present disclosure provides a detailed description of the vehicle cutting-in detection method based on monocular images provided in the embodiments of this disclosure.

[0038] Figure 1 A schematic flowchart of a vehicle cutting in front method based on monocular images provided in an embodiment of this disclosure is shown.

[0039] In this embodiment of the disclosure, the forward vehicle cut-in detection method based on monocular images can be executed by an electronic device. This electronic device may include, but is not limited to, mobile terminals such as mobile phones, in-vehicle devices, vehicle controllers, tablets, and wearable devices.

[0040] like Figure 1 As shown, the method for detecting a vehicle cutting in front based on a monocular image may include the following steps.

[0041] S110. Acquire consecutive multi-frame target detection images, the target detection images including the detection boxes of the vehicles in front located in adjacent lanes.

[0042] In this embodiment of the disclosure, the electronic device may first acquire multiple consecutive frames of target detection images.

[0043] Optionally, the object detection image can be a detection image obtained by performing object detection (OD) on the acquired road image. Object detection can involve identifying vehicles in the acquired road image and displaying detection boxes. The object detection image may include detection boxes for vehicles located in adjacent lanes.

[0044] For example, when a vehicle equipped with electronic devices such as in-vehicle equipment is driving on a road, it can acquire multiple consecutive frames of road images ahead through a monocular camera, and then the computing unit in the in-vehicle equipment can perform target detection (OD) and lane detection on these multiple frames of road images to obtain corresponding consecutive multiple frames of target detection images. These multiple frames of target detection images may include the detection boxes of the vehicles in front located in adjacent lanes, and may also include the corresponding lane lines, etc.

[0045] Figure 2 A schematic diagram of the structure of a target detection image provided in an embodiment of this disclosure is shown.

[0046] like Figure 2 As shown, it includes the detection box corresponding to the preceding vehicle 201 in the adjacent lane 202, and lane 203 is the driving lane for the vehicle, and also includes the corresponding lane lines.

[0047] S120. Obtain the first ranging point on the detection box of the front vehicle in each frame of the target detection image. The first ranging point is the center point of the bottom edge of the rear end of the front vehicle.

[0048] In this embodiment of the disclosure, after acquiring multiple consecutive target detection images, the electronic device can acquire the first ranging point on the detection frame of the vehicle in front in each target detection image.

[0049] Optionally, the first distance measuring point can be a reference point used to detect the distance between the vehicle in front and the vehicle itself. Specifically, the first distance measuring point can be the center point of the rear bottom edge of the vehicle in front.

[0050] For example, after obtaining multiple consecutive frames of target detection images, the electronic device can, for each frame of the target detection image, obtain the first ranging point on the detection box of the preceding vehicle in the target detection image using algorithms such as BEV algorithm or IPM algorithm. Specific implementation methods are described below.

[0051] S130. For each frame of target detection image, obtain the relative motion parameters between the vehicle in front and the vehicle itself based on the first ranging point. The relative motion parameters include the relative lateral distance and relative lateral velocity of the vehicle in front.

[0052] In this embodiment of the disclosure, after acquiring the first ranging point on the detection frame of the preceding vehicle in each frame of the target detection image, the electronic device can obtain the relative motion parameters between the preceding vehicle and the vehicle itself based on the corresponding first ranging point for each frame of the target detection image.

[0053] Optionally, the relative motion parameters can be parameters used to characterize the motion state of the preceding vehicle relative to the following vehicle. These relative motion parameters may include relative lateral distance and relative lateral velocity.

[0054] Optionally, the relative lateral distance can be the lateral distance of the preceding vehicle on the horizontal axis of a spatial coordinate system centered on the vehicle itself; the relative lateral velocity can be the lateral velocity of the preceding vehicle on the horizontal axis of a spatial coordinate system centered on the vehicle itself.

[0055] Specifically, for each frame of target detection image, the electronic device can smooth the first ranging point by means of filtering, Kalman filtering, etc., to obtain the relative lateral distance between the preceding vehicle and the following vehicle. Then, based on the relative lateral distance and the image time, the corresponding relative lateral velocity is calculated, thereby obtaining the corresponding relative motion parameters.

[0056] S140. If it is determined that the vehicle in front has a tendency to approach based on the relative lateral distance and relative lateral velocity corresponding to the target detection images in multiple consecutive frames, the first distance measurement point is updated by calculating the intersection of three planes for every two consecutive target detection images in the multiple frames, and the vehicle in front has a tendency to cut in is determined based on the updated first distance measurement point.

[0057] In this embodiment of the disclosure, after obtaining the relative lateral distance and relative lateral velocity of the vehicle in front in each frame of target detection image, the electronic device can determine that the vehicle in front has a tendency to approach based on the relative lateral distance and relative lateral velocity corresponding to multiple consecutive frames of target detection images. Then, it can calculate the intersection of three planes for every two consecutive frames of target detection images in the multiple frames of target detection images to obtain an updated first ranging point, and determine whether the vehicle in front has a tendency to cut in based on the updated first ranging point.

[0058] Optionally, the tendency for the preceding vehicle to approach can be defined as the tendency for the preceding vehicle to gradually approach the following vehicle.

[0059] Alternatively, the intersection of the three planes can be used to calculate the new first distance measurement point.

[0060] Optionally, the tendency of the preceding vehicle to cut in can be the tendency of the preceding vehicle to want to enter the driving lane from the adjacent lane.

[0061] Specifically, after obtaining the relative lateral distance and relative lateral velocity of the vehicle in front in each frame of target detection image, the electronic device can calculate the intersection of three planes in every two consecutive frames of target detection images based on the relative lateral distance and relative lateral velocity of the vehicle in front in multiple consecutive frames of target detection images, and if it is determined that the vehicle in front has a tendency to approach, thereby obtaining the updated first ranging point. Then, it determines whether the vehicle in front has a tendency to cut in based on the updated first ranging point.

[0062] Therefore, in this embodiment, it is possible to acquire a series of multi-frame target detection images including the detection frame of the vehicle in the adjacent lane. Then, a first ranging point on the detection frame of the vehicle in each frame is acquired. This first ranging point is the center point of the rear bottom edge of the vehicle. For each frame, relative motion parameters, including relative lateral distance and relative lateral velocity, are obtained between the vehicle and the vehicle based on the first ranging point. Finally, if it is determined that the vehicle is approaching based on the relative lateral distance and relative lateral velocity corresponding to the series of multi-frame target detection images, three-dimensional motion is performed on every two consecutive frames of target detection images. The updated first distance measurement point is obtained by calculating the intersection point of the plane. Based on the updated first distance measurement point, it is determined whether the vehicle in front has a tendency to cut in. Thus, the relative motion parameters, namely the relative lateral distance and the relative lateral velocity, can be obtained first through the first distance measurement point. Then, if it is determined that the vehicle in front has a tendency to approach based on the relative lateral distance and the relative lateral velocity, the updated first distance measurement point is obtained. Finally, the updated first distance measurement point is used to determine whether the vehicle in front has a tendency to cut in. In this way, the distance measurement and detection of the vehicle in front can be performed by using the first distance measurement point and the updated first distance measurement point, thereby improving the accuracy of distance measurement and detection, reducing distance measurement and detection errors, and improving vehicle driving safety.

[0063] Optionally, S120 may specifically include: calculating the target endpoint coordinates on the detection box of the preceding vehicle in each frame of the target detection image, the target endpoint coordinates including the first endpoint coordinates and the second endpoint coordinates, the first endpoint coordinates being the endpoint coordinates of the bottom edge of the detection box of the preceding vehicle that is far from the target lane line, and the second endpoint coordinates being the endpoint coordinates of the bottom edge of the detection box of the preceding vehicle that is close to the target lane line, the target lane line being the lane line between the vehicle and the preceding vehicle; calculating the difference in the abscissas of the first endpoint coordinates and the second endpoint coordinates to obtain the first width of the detection box of the preceding vehicle; performing similarity calculations using the first endpoint coordinates, the second endpoint coordinates, the first width, and the second width to obtain the third endpoint coordinates on the detection box of the preceding vehicle, the second width being the actual width of the preceding vehicle; and obtaining the center coordinates of the third endpoint coordinates and the first endpoint coordinates to obtain the coordinates of the first ranging point.

[0064] In this embodiment of the disclosure, after acquiring multiple frames of target detection images, the electronic device can first calculate the target endpoint coordinates on the detection frame of the vehicle in front in each frame of target detection images.

[0065] Optionally, the target endpoint coordinates may include the first endpoint coordinates and the second endpoint coordinates.

[0066] Optionally, the coordinates of the first endpoint can be the coordinates of the endpoint of the detection frame of the preceding vehicle that is furthest from the target lane line.

[0067] Optionally, the coordinates of the second endpoint can be the coordinates of the endpoint of the detection frame of the preceding vehicle that is closest to the target lane line.

[0068] Optionally, the target lane line can be the lane line between the vehicle and the vehicle in front.

[0069] Continue to refer to Figure 2 The target lane line 204 can be the lane line between the preceding vehicle 201 and the vehicle itself. In the detection box corresponding to the preceding vehicle 201, the first endpoint coordinate can be the coordinate of point C, and the second endpoint coordinate can be the coordinate of point D.

[0070] Specifically, the electronic device can first determine the direction of the lane line, and then use the BEV algorithm to calculate the coordinates of the first endpoint and the second endpoint of the detection box of the preceding vehicle in each frame of the target detection image. The first endpoint coordinate can be the coordinate of the bottom edge of the detection box of the preceding vehicle that is far away from the target lane line, and the second endpoint coordinate can be the coordinate of the bottom edge of the detection box of the preceding vehicle that is close to the target lane line.

[0071] Furthermore, after obtaining the coordinates of the first endpoint and the second endpoint, the electronic device can calculate the difference in the horizontal coordinates of the first endpoint and the second endpoint to obtain the first width of the detection frame of the vehicle in front.

[0072] Specifically, the electronic device can subtract the abscissas of the first endpoint coordinates and the second endpoint coordinates to obtain the distance between the first endpoint coordinates and the second endpoint coordinates, which is the first width corresponding to the detection frame of the vehicle in front.

[0073] Furthermore, the electronic device can perform similarity calculations using the first endpoint coordinates, the second endpoint coordinates, the first width, and the second width to obtain the third endpoint coordinates on the detection frame of the preceding vehicle.

[0074] Optionally, the second width can be the actual width used to represent the vehicle in front.

[0075] Alternatively, similarity calculation can be performed based on similarity relations.

[0076] Continue to refer to Figure 2 Let the first width be d1 and the second width be d2. Using the coordinates of point C (the first endpoint coordinates), the coordinates of point D (the second endpoint coordinates), d1, and d2, the corresponding similarity relationship can be expressed as CF / CD = d2 / d1 => CF = d2 / d1 * CD. Thus, the coordinates of point F in the detection box, i.e., the coordinates of the third endpoint, can be obtained.

[0077] Furthermore, after obtaining the coordinates of the third endpoint, the electronic device can obtain the center coordinates of the third endpoint coordinates and the first endpoint coordinates, thereby obtaining the coordinates of the first ranging point.

[0078] Continue to refer to Figure 2After obtaining the coordinates of the third endpoint, point F, the electronic device can calculate the coordinates of the center point N, i.e., the coordinates of the first ranging point, based on the coordinates of point F and point C. Compared to the center point M of points C and D, point N is more accurate, resulting in smaller ranging detection errors, as point N is closer to the actual center point of the vehicle's rear bottom edge.

[0079] Furthermore, after acquiring the first ranging point on the detection frame of the preceding vehicle in each frame of the target detection image, the electronic device can obtain the relative motion parameters between the preceding vehicle and the vehicle itself based on the corresponding first ranging point for each frame of the target detection image, that is, obtain the relative lateral distance and the relative lateral velocity.

[0080] Therefore, in this embodiment of the present disclosure, the accuracy of ranging detection can be improved, the ranging detection error can be reduced, and the vehicle driving safety can be improved.

[0081] In some embodiments, the electronic device can identify the vehicle type of the preceding vehicle in the target detection image; and determine the corresponding second width based on the vehicle type of the preceding vehicle.

[0082] In this embodiment of the disclosure, the electronic device can identify the vehicle type of the vehicle in front of it by performing image recognition on the target detection image, such as the vehicle being a bus, a sedan, a truck, etc.

[0083] Furthermore, the electronic device can determine the corresponding second width, i.e. the corresponding actual width, of the vehicle in front based on the identified vehicle type. For example, the second width can be 1.6m when the vehicle in front is a sedan, and 5m when the vehicle in front is a bus, etc. There are no restrictions here.

[0084] Furthermore, the electronic device can perform similarity calculations using the first endpoint coordinates, the second endpoint coordinates, the first width, and the second width to obtain the third endpoint coordinates on the detection frame of the preceding vehicle.

[0085] Optionally, the relative motion parameters may also include the relative longitudinal distance and relative longitudinal speed of the preceding vehicle.

[0086] Optionally, the relative longitudinal distance can be the longitudinal distance of the preceding vehicle on the vertical axis of a spatial coordinate system centered on the vehicle itself; the relative longitudinal speed can be the longitudinal speed of the preceding vehicle on the horizontal axis of a spatial coordinate system centered on the vehicle itself.

[0087] In this process, the electronic device can smooth the first ranging point for each frame of target detection image using filtering methods such as mean and Kalman filtering to obtain the relative longitudinal distance between the preceding vehicle and the vehicle itself. Then, based on the relative longitudinal distance and the image time, it can calculate the corresponding relative longitudinal velocity, thereby obtaining the corresponding relative motion parameters.

[0088] In some embodiments of this disclosure, the electronic device can also obtain the relative longitudinal distance and relative longitudinal speed between the vehicle in front and the vehicle itself. When the relative longitudinal distance and relative longitudinal speed between the vehicle in front are large, the actual distance between the vehicle in front and the vehicle itself is also large. In this case, the electronic device does not need to perform further distance measurement and detection, and the vehicle in front poses less of a threat to the vehicle itself.

[0089] In some other embodiments of this disclosure, when the relative longitudinal distance and relative longitudinal speed of the preceding vehicle are small, the electronic device needs to further determine the relative lateral distance and relative lateral speed of the preceding vehicle. The electronic device determines that the preceding vehicle has a tendency to approach based on the relative lateral distance and relative lateral speed corresponding to multiple consecutive target detection images.

[0090] Optionally, S140 may specifically include: if the third endpoint coordinate does not exceed the target lane line, and the relative lateral distance gradually decreases while the relative lateral speed is not zero, then it is determined that the vehicle in front has a tendency to approach, and the third endpoint coordinate is the coordinate of the bottom end of the rear end of the vehicle in front; if the third endpoint coordinate exceeds the target lane line, it is determined that the vehicle in front has a tendency to approach.

[0091] In some embodiments of this disclosure, after obtaining the relative lateral distance and relative lateral velocity of the vehicle ahead in each frame of target detection image, if the relative lateral distance gradually decreases and the relative lateral velocity is not zero, the electronic device determines that the vehicle ahead has a tendency to approach, provided that the coordinates of the third endpoint do not exceed the target lane line.

[0092] Optionally, the coordinates of the third endpoint can be the coordinates of the bottom edge of the rear end of the front vehicle.

[0093] Continue to refer to Figure 2 The coordinates of the third endpoint can be the coordinates of point F. The electronic device needs to first determine that the coordinates of point F have not exceeded the target lane line 204. Then, in multiple consecutive target detection images, the relative lateral distance of the vehicle 201 in front is gradually decreasing. If the relative lateral distance of the vehicle 201 in front is gradually decreasing in three consecutive target detection images and the relative lateral speed of the vehicle 201 in front is not zero, the electronic device can determine that the vehicle 201 in front has a tendency to approach.

[0094] Specifically, when the electronic device determines that the third endpoint coordinate of the detection frame of the preceding vehicle has not exceeded the target lane line, it can continue to judge the relative lateral distance and relative lateral velocity of the preceding vehicle. If the relative lateral distance of the preceding vehicle gradually decreases for several consecutive frames and the relative lateral velocity of the preceding vehicle is not zero, it can be determined that the preceding vehicle has a tendency to approach. Similarly, if the relative lateral distance of the preceding vehicle does not gradually decrease for several consecutive frames, or the relative lateral velocity of the preceding vehicle is zero, it can be determined that the preceding vehicle does not have a tendency to approach.

[0095] In other embodiments of this disclosure, after obtaining the relative lateral distance and relative lateral velocity of the vehicle ahead in each frame of target detection image, the electronic device determines that the vehicle ahead has a tendency to approach if it determines that the coordinates of the third endpoint exceed the target lane line.

[0096] Continue to refer to Figure 2 The coordinates of the third endpoint can be the coordinates of point F. If the electronic device determines that the coordinates of point F exceed the target lane line 204, the electronic device can determine that the vehicle in front is approaching.

[0097] Specifically, when the electronic device determines that the coordinates of the third endpoint on the detection frame of the vehicle in front have exceeded the target lane line, the electronic device no longer needs to determine the relative lateral distance and relative lateral speed of the vehicle in front, and can then determine that the vehicle in front has a tendency to approach.

[0098] Therefore, in this embodiment of the present disclosure, the electronic device can comprehensively determine whether the vehicle in front is approaching by using the coordinates of the third endpoint, the relative lateral distance, and the relative lateral speed, which can improve the accuracy of distance measurement and detection, reduce distance measurement and detection errors, and improve vehicle driving safety.

[0099] In some examples, the electronic equipment needs to update the first ranging point after determining that the vehicle in front is approaching, which will be explained in detail below.

[0100] Optionally, S140 may further include: obtaining a depth image based on a monocular depth estimation algorithm; acquiring the side plane equation and rear plane equation of the vehicle in front in the depth image; calculating the intersection of the side plane equation, rear plane equation and ground plane equation to obtain the updated first ranging point.

[0101] In this embodiment of the disclosure, after determining that the vehicle in front is approaching, the electronic device can obtain a depth image of the target detection image based on a monocular depth estimation algorithm.

[0102] For example, monocular depth estimation algorithms can be MonoDepth2, DPT, Midas algorithm, etc., and no specific limitation is made here.

[0103] Specifically, for every two consecutive frames of target detection images, the electronic device generates a corresponding depth image based on a monocular depth estimation algorithm, and fits the corresponding ground plane equation to the point cloud corresponding to the depth image.

[0104] For example, the equation for the ground plane can be expressed as ax + by + cz + d = 0.

[0105] Furthermore, the electronic device calculates the camera height in the depth image.

[0106] The corresponding calculation formula is:

[0107]

[0108] Where h1 represents the camera height.

[0109] Optionally, the electronic device can calculate the corresponding point cloud scale using prior camera height information h0 and camera height h1. The point cloud scale represents the proportional relationship between length and true length in the depth image.

[0110] Furthermore, the electronic device can obtain the side plane equations and rear plane equations of the vehicle in front from the depth image based on the corresponding point cloud. That is, it can obtain... Figure 2 The equations corresponding to the ADFG and BCFG planes are given. The equations for the lateral and rear planes are similar to those for the ground plane, and are not specified here.

[0111] Furthermore, the electronic device can calculate the intersection of the side plane equation, the rear plane equation, and the ground plane equation to obtain an updated first ranging point. Compared with the first ranging point, this updated first ranging point does not use the actual width information of the vehicle in front, i.e., the second width. Therefore, the updated first ranging point is more accurate, thus making the ranging detection more accurate.

[0112] Optionally, S140 may further include: determining that the preceding vehicle has a tendency to cut in if the updated first distance measurement point exceeds the target lane line.

[0113] In this embodiment of the disclosure, after obtaining the updated first distance measurement point, if the updated first distance measurement point exceeds the target lane line, the electronic device can determine that the vehicle in front has a tendency to cut in.

[0114] Optionally, after determining that the vehicle in front is inclined to cut in, the electronic device can take corresponding actions in a timely manner, such as controlling the vehicle's braking system to decelerate, or alerting the driver through sound and lights, etc., which are not limited here.

[0115] Figure 3 A schematic diagram of a vehicle cutting in front based on a monocular image provided in an embodiment of this disclosure is shown.

[0116] like Figure 3 As shown, the monocular image-based vehicle cutting-in detection device 300 may include a first acquisition module 310, a second acquisition module 320, a first processing module 330, and a second processing module 340.

[0117] The first acquisition module 310 can be used to acquire multiple consecutive target detection images, which include the detection box of the vehicle in front located in the adjacent lane.

[0118] The second acquisition module 320 can be used to acquire the first ranging point on the detection box of the front vehicle in each frame of the target detection image. The first ranging point is the center point of the bottom edge of the rear end of the front vehicle.

[0119] The first processing module 330 can be used to obtain the relative motion parameters between the vehicle in front and the vehicle itself based on the first ranging point for each frame of target detection image. The relative motion parameters include the relative lateral distance and relative lateral speed of the vehicle in front.

[0120] The second processing module 340 can be used to calculate the intersection of three planes in every two consecutive target detection images in the multi-frame target detection images to obtain the updated first ranging point when it is determined that the vehicle in front has a tendency to approach based on the relative lateral distance and relative lateral velocity corresponding to the target detection images in consecutive frames, and to determine whether the vehicle in front has a tendency to cut in based on the updated first ranging point.

[0121] Therefore, in this embodiment, it is possible to acquire a series of multi-frame target detection images including the detection frame of the vehicle in the adjacent lane. Then, a first ranging point on the detection frame of the vehicle in each frame is acquired. This first ranging point is the center point of the rear bottom edge of the vehicle. For each frame, relative motion parameters, including relative lateral distance and relative lateral velocity, are obtained between the vehicle and the vehicle based on the first ranging point. Finally, if it is determined that the vehicle is approaching based on the relative lateral distance and relative lateral velocity corresponding to the series of multi-frame target detection images, three-dimensional motion is performed on every two consecutive frames of target detection images. The updated first distance measurement point is obtained by calculating the intersection point of the plane. Based on the updated first distance measurement point, it is determined whether the vehicle in front has a tendency to cut in. Thus, the relative motion parameters, namely the relative lateral distance and the relative lateral velocity, can be obtained first through the first distance measurement point. Then, if it is determined that the vehicle in front has a tendency to approach based on the relative lateral distance and the relative lateral velocity, the updated first distance measurement point is obtained. Finally, the updated first distance measurement point is used to determine whether the vehicle in front has a tendency to cut in. In this way, the distance measurement and detection of the vehicle in front can be performed by using the first distance measurement point and the updated first distance measurement point, thereby improving the accuracy of distance measurement and detection, reducing distance measurement and detection errors, and improving vehicle driving safety.

[0122] In some embodiments of this disclosure, the second acquisition module 320 may include a first calculation unit, a second calculation unit, a third calculation unit, and a first acquisition unit.

[0123] The first computing unit can be used to calculate the target endpoint coordinates on the detection box of the preceding vehicle in each frame of the target detection image. The target endpoint coordinates include the first endpoint coordinates and the second endpoint coordinates. The first endpoint coordinates are the endpoint coordinates of the bottom edge of the detection box of the preceding vehicle that is far away from the target lane line, and the second endpoint coordinates are the endpoint coordinates of the bottom edge of the detection box of the preceding vehicle that is close to the target lane line. The target lane line is the lane line between the vehicle and the preceding vehicle.

[0124] The second calculation unit can be used to calculate the difference in the horizontal coordinates of the first endpoint coordinates and the second endpoint coordinates to obtain the first width of the detection frame of the vehicle in front.

[0125] The third calculation unit can be used to perform similarity calculations using the first endpoint coordinates, the second endpoint coordinates, the first width, and the second width to obtain the third endpoint coordinates on the detection frame of the preceding vehicle, where the second width is the actual width of the preceding vehicle.

[0126] The first acquisition unit can be used to obtain the center coordinates of the third endpoint coordinates and the first endpoint coordinates to obtain the coordinates of the first distance measuring point.

[0127] In some embodiments of this disclosure, the monocular image-based vehicle cutting-in detection device 300 may include a vehicle recognition module and a width determination module.

[0128] This vehicle recognition module can be used to identify the vehicle type of the vehicle in front of the target detection image.

[0129] This width determination module can be used to determine the corresponding second width based on the vehicle type of the vehicle in front.

[0130] In some embodiments of this disclosure, the relative motion parameters may further include the relative longitudinal distance and relative longitudinal speed of the preceding vehicle.

[0131] In some embodiments of this disclosure, the second processing module 340 may include a first determining unit and a second determining unit.

[0132] The first determining unit can be used to determine that the vehicle in front has a tendency to approach if the relative lateral distance gradually decreases and the relative lateral speed is not zero, when the coordinates of the third endpoint do not exceed the target lane line. The coordinates of the third endpoint are the coordinates of the bottom edge of the rear end of the vehicle in front.

[0133] The second determining unit can be used to determine that the vehicle in front has a tendency to approach when the coordinates of the third endpoint exceed the target lane line.

[0134] In some embodiments of this disclosure, the second processing module 340 may further include a depth estimation unit, a second acquisition unit, and a fourth calculation unit.

[0135] This depth estimation unit can be used to obtain depth images based on monocular depth estimation algorithms.

[0136] The second acquisition unit can be used to acquire the side plane equation and rear plane equation of the vehicle in front in the depth image.

[0137] The fourth calculation unit can be used to calculate the intersection of the side plane equation, the back plane equation, and the ground plane equation to obtain the updated first distance measurement point.

[0138] In some embodiments of this disclosure, the second processing module 340 may further include a third determining unit.

[0139] The third determining unit can be used to determine that the vehicle ahead has a tendency to cut in if the updated first distance measurement point exceeds the target lane line.

[0140] It should be noted that, Figure 3 The monocular image-based vehicle incursion detection device 300 shown can perform... Figure 1 The various steps in the method embodiment shown are implemented. Figure 1 The processes and effects in the method embodiments shown are not described in detail here.

[0141] Figure 4 A schematic diagram of the structure of a vehicle cutting in based on monocular images provided in an embodiment of this disclosure is shown.

[0142] In some embodiments of this disclosure, Figure 4 The monocular image-based vehicle cut-in detection device shown can be any electronic device that a user wants to perform monocular image-based vehicle cut-in detection. Specifically, the electronic device can include, but is not limited to, mobile terminals such as mobile phones, in-vehicle devices, vehicle controllers, tablets, wearable devices, and smart home devices.

[0143] like Figure 4 As shown, the monocular image-based vehicle cutting-in detection device may include a processor 401 and a memory 402 storing computer program instructions.

[0144] Specifically, the processor 401 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0145] Memory 402 may include a large-capacity storage for information or instructions. For example, and not limitingly, memory 402 may include a hard disk drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 402 may include removable or non-removable (or fixed) media. Where appropriate, memory 402 may be internal or external to the integrated gateway device. In a particular embodiment, memory 402 is a non-volatile solid-state memory. In a particular embodiment, memory 402 includes read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (Electrically Programmable ROM, EPROM), an electrically erasable programmable PROM (EEPROM), an electrically alterable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0146] The processor 401 reads and executes computer program instructions stored in the memory 402 to perform the steps of the monocular image-based vehicle cutting-in detection method provided in this embodiment of the present disclosure.

[0147] In one example, the monocular image-based vehicle cut-in detection device may further include a transceiver 403 and a bus 404. Wherein, as Figure 4 As shown, the processor 401, memory 402 and transceiver 403 are connected via bus 404 and communicate with each other.

[0148] Bus 404 includes hardware, software, or both. For example, and not limitingly, a bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industrial Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 404 may include one or more buses. Although specific buses are described and illustrated in the embodiments of this application, this application considers any suitable bus or interconnection.

[0149] This disclosure also provides a computer-readable storage medium that can store a computer program. When the computer program is executed by a processor, the processor enables the processor to implement the monocular image-based vehicle cutting-in detection method provided in this disclosure.

[0150] The aforementioned storage medium may, for example, include a memory 402 containing computer program instructions, which can be executed by the processor 401 of the monocular image-based vehicle entry detection device to complete the monocular image-based vehicle entry detection method provided in this embodiment. Optionally, the storage medium may be a non-transitory computer-readable storage medium, such as a ROM, random access memory (RAM), compact disc ROM (CD-ROM), magnetic tape, floppy disk, or optical monocular image-based vehicle entry detection device.

[0151] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the term "comprising" is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0152] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting vehicle cutting in front based on monocular images, characterized in that, include: Acquire consecutive multi-frame target detection images, wherein the target detection images include detection boxes of the preceding vehicle located in the adjacent lane; Obtain the first ranging point on the detection frame of the front vehicle in each frame of the target detection image, where the first ranging point is the center point of the bottom edge of the rear end of the front vehicle; For each frame of the target detection image, the relative motion parameters between the vehicle in front and the vehicle itself are obtained based on the first ranging point. The relative motion parameters include the relative lateral distance and relative lateral velocity of the vehicle in front. If it is determined that the vehicle in front has a tendency to approach based on the relative lateral distance and relative lateral velocity corresponding to multiple consecutive frames of the target detection images, the first distance measurement point is updated by calculating the intersection of three planes for every two consecutive frames of the target detection images, and the vehicle in front has a tendency to cut in is determined based on the updated first distance measurement point. The step of obtaining the first ranging point on the detection frame of the preceding vehicle in each frame of the target detection image includes: Calculate the target endpoint coordinates on the detection box of the preceding vehicle in each frame of the target detection image. The target endpoint coordinates include a first endpoint coordinate and a second endpoint coordinate. The first endpoint coordinate is the endpoint coordinate of the bottom edge of the detection box of the preceding vehicle that is far away from the target lane line. The second endpoint coordinate is the endpoint coordinate of the bottom edge of the detection box of the preceding vehicle that is close to the target lane line. The target lane line is the lane line between the vehicle and the preceding vehicle. Calculate the difference in the abscissas between the coordinates of the first endpoint and the coordinates of the second endpoint to obtain the first width of the detection frame of the vehicle in front; By performing similarity calculations using the first endpoint coordinates, the second endpoint coordinates, the first width, and the second width, the third endpoint coordinates on the detection frame of the preceding vehicle are obtained, and the second width is the actual width of the preceding vehicle. The coordinates of the first distance measuring point are obtained by finding the center coordinates of the third endpoint coordinates and the first endpoint coordinates. The step of determining that the vehicle in front has a tendency to approach based on the relative lateral distance and relative lateral velocity corresponding to multiple consecutive frames of the target detection images includes: If the relative lateral distance gradually decreases and the relative lateral speed is not zero when the third endpoint coordinate does not exceed the target lane line, it is determined that the vehicle in front has a tendency to approach, and the third endpoint coordinate is the bottom endpoint coordinate of the rear end of the vehicle in front. If the coordinates of the third endpoint exceed the target lane line, it is determined that the vehicle in front has a tendency to approach.

2. The method according to claim 1, characterized in that, The method further includes: Identify the vehicle type of the preceding vehicle in the target detection image; The corresponding second width is determined based on the vehicle type of the preceding vehicle.

3. The method according to claim 1, characterized in that, The relative motion parameters also include the relative longitudinal distance and relative longitudinal speed of the preceding vehicle.

4. The method according to claim 1, characterized in that, The step of calculating the intersection of three planes in every two consecutive target detection images from multiple frames of the target detection images to obtain the updated first ranging point includes: A depth image is obtained based on a monocular depth estimation algorithm; Obtain the side plane equation and rear plane equation of the front vehicle in the depth image; The intersection of the side plane equation, the back plane equation, and the ground plane equation is calculated to obtain the updated first ranging point.

5. The method according to claim 1, characterized in that, Determining whether the preceding vehicle has a cutting-in tendency based on the updated first ranging point includes: If the updated first distance measurement point exceeds the target lane line, it is determined that the vehicle in front has a tendency to cut in.

6. A device for detecting vehicle intrusion based on monocular images, characterized in that, include: The first acquisition module is used to acquire multiple consecutive target detection images, wherein the target detection images include the detection box of the vehicle in front located in the adjacent lane; The second acquisition module is used to acquire the first ranging point on the detection frame of the front vehicle in each frame of the target detection image, wherein the first ranging point is the center point of the bottom edge of the rear end of the front vehicle; The first processing module is used to obtain the relative motion parameters between the vehicle in front and the vehicle itself based on the first ranging point for each frame of the target detection image. The relative motion parameters include the relative lateral distance and relative lateral velocity of the vehicle in front. The second processing module is used to calculate the intersection of three planes in every two consecutive target detection images in the multiple target detection images to obtain an updated first ranging point when it is determined that the vehicle in front has a tendency to approach based on the relative lateral distance and relative lateral velocity corresponding to the target detection images in multiple consecutive frames, and to determine whether the vehicle in front has a tendency to cut in based on the updated first ranging point. The second acquisition module includes a first calculation unit, a second calculation unit, a third calculation unit, and a first acquisition unit. The first calculation unit is used to calculate the target endpoint coordinates on the detection box of the preceding vehicle in each frame of target detection image. The target endpoint coordinates include the first endpoint coordinates and the second endpoint coordinates. The first endpoint coordinates are the endpoint coordinates of the bottom edge of the detection box of the preceding vehicle that is far away from the target lane line, and the second endpoint coordinates are the endpoint coordinates of the bottom edge of the detection box of the preceding vehicle that is close to the target lane line. The target lane line is the lane line between the vehicle and the preceding vehicle. The second calculation unit is used to calculate the difference in the horizontal coordinates of the first endpoint coordinates and the second endpoint coordinates to obtain the first width of the detection frame of the vehicle in front. The third calculation unit is used to perform similarity calculations using the first endpoint coordinates, the second endpoint coordinates, the first width, and the second width to obtain the third endpoint coordinates on the detection frame of the preceding vehicle, where the second width is the actual width of the preceding vehicle. The first acquisition unit can be used to acquire the center coordinates of the third endpoint coordinates and the first endpoint coordinates to obtain the coordinates of the first ranging point; The second processing module includes a first determining unit and a second determining unit; The first determining unit is used to determine that the vehicle in front has a tendency to approach if the relative lateral distance gradually decreases and the relative lateral speed is not zero when the third endpoint coordinate does not exceed the target lane line. The third endpoint coordinate is the coordinate of the bottom edge of the rear end of the vehicle in front. The second determining unit is used to determine that the vehicle in front has a tendency to approach when the coordinates of the third endpoint exceed the target lane line.

7. A vehicle cutting-in detection device based on monocular images, characterized in that, include: processor; Memory, used to store executable instructions; The processor is configured to read the executable instructions from the memory and execute the executable instructions to implement the vehicle cutting-in detection method based on monocular images as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, causes the processor to implement the vehicle cutting-in detection method based on monocular images as described in any one of claims 1-5.

Images