A grid-based railing detection method based on BSD radar

Abstract

The present invention relates to a grid-based railing detection method based on a BSD radar, comprising: S10. Establishing a vehicle body coordinate system, and dividing the railing area into multiple grid areas at a fixed interval in the vehicle body coordinate system; S20. Screening and processing the stationary points detected by each frame of the BSD radar and storing them in the divided grid areas, so that at most one stationary point is retained in each grid area; S30. Performing motion compensation on the stationary points in each grid area according to the vehicle speed of the vehicle itself and updating the grid areas; S40. Performing a third-order fitting on the stationary points in all grid areas over a period of time to obtain a railing fitting curve to determine whether the railing is valid; S50. If the railing is a valid railing, labeling the stationary points not placed in the grid areas. By dividing the railing area into multiple grid areas at a fixed interval and storing the stationary points detected by multiple frames of the BSD radar in multiple grid areas respectively, the present invention screens the stored data in the grid areas, reducing the amount of computation and storage.

Description

Technical Field

[0001] The present invention relates to the technical field of railing recognition, and particularly to a grid-based railing detection method based on a BSD radar. Background Art

[0002] Affected by the detection angle accuracy and speed accuracy of the BSD millimeter-wave radar, and complex scenarios such as railings and isolation green belts often accompany the actual vehicle driving, it is easy to recognize false targets, resulting in false alarms of the BSD. Therefore, the recognition of the railing scenario is particularly important.

[0003] Currently, the technologies for the BSD millimeter-wave radar to recognize railings include the recognition by the histogram method, the recognition by the Hough transform method, and the recognition by the target point clustering method. Each method has its own disadvantages. The recognition by the histogram method will cause a large error between the recognized position of the railing and the actual position due to the segmentation at a fixed distance; the recognition by the Hough transform method is relatively complex in calculation and has low efficiency; the recognition by the target point clustering method is not obvious when the number of detected targets is small, and the set parameters of the clustering gate need to be adjusted at different distances and scenarios, and a large amount of actual vehicle data is required for verification.

[0004] With the improvement of the resolution of the BSD millimeter-wave radar, the number of point clouds for railing detection increases. The number of point clouds in a single frame is limited, and affected by the detection accuracy, it is not possible to distinguish well between railings and targets. The number of point clouds stored in multiple frames is too large, and the amount of calculation and storage often encounter engineering problems where the calculation amount is too large to be realized or the performance degrades.

[0005] Therefore, how to better distinguish the actual application scenario of the railing while reducing the calculation amount of the BSD millimeter-wave radar will be a problem to be solved. Summary of the Invention

[0006] In order to overcome the problems of large error, low efficiency, and large calculation amount existing in the above-mentioned existing technologies for the BSD millimeter-wave radar to recognize railings, the present invention provides a grid-based railing detection method based on a BSD radar.

[0007] To solve the above technical problems, the technical solution of the present invention is as follows:

[0008] A grid-based railing detection method based on a BSD radar, comprising the steps of:

[0009] S10. Establish a vehicle body coordinate system, and divide the railing area into multiple grid areas at a fixed interval in the vehicle body coordinate system;

[0010] S20. Screen and process the stationary points detected by the BSD radar in each frame, and store them in the divided grid areas, so that at most one stationary point is retained in each grid area;

[0011] S30. Perform motion compensation on the stationary points in each grid area according to the vehicle speed, and update the grid area;

[0012] S40. Perform third-order fitting on the stationary points in all grid areas within a period of time to obtain the railing fitting curve, and determine whether the railing is valid according to the railing fitting curve;

[0013] S50. If the railing is a valid railing, label the stationary points that have not been placed in the grid area.

[0014] Further, as a preferred technical solution, step S20 specifically includes:

[0015] S201. Screen the stationary points detected by the BSD radar in each frame, and store the stationary points in the divided grid areas according to their coordinate positions;

[0016] S202. Perform edge extraction on the multiple stationary points placed in each grid area so that each grid area retains at most one stationary point.

[0017] Further, as a preferred technical solution, in step S202, edge extraction needs to be performed on the stationary points stored in the grid area in each frame to ensure that at most one stationary point is stored in each grid area.

[0018] Further, as a preferred technical solution, step S30 specifically includes:

[0019] S301. According to the relative position coordinates of the stationary points in each grid area in the vehicle body coordinate system, the vehicle speed of the host vehicle, and the time interval between two adjacent frames, perform position compensation on the stationary points in each grid area, update the coordinate positions of the stationary points, and update the grid area.

[0020] Further, as a preferred technical solution, step S30 further includes:

[0021] S302. According to the distribution of the stationary points in the updated grid area, perform edge extraction on the grid areas with multiple stationary points so that each grid area retains at most one stationary point, and update the grid area.

[0022] Further, as a preferred technical solution, step S40 specifically includes:

[0023] S401. According to the coordinate positions of the stationary points in the grid area within a period of time, use polynomial fitting to obtain the railing fitting coefficients;

[0024] S402. Obtain the railing fitting curve and the railing fitting deviation according to the railing fitting coefficients;

[0025] S403. When the railing fitting deviation is less than the preset fitting deviation value, it is determined that the railing fitting is effective.

[0026] Further, as a preferred technical solution, the preset fitting deviation value is the maximum value allowed for the railing fitting deviation set according to the BSD radar road test data.

[0027] Further, as a preferred technical solution, step S50 specifically includes:

[0028] If the railing is an effective railing, according to the railing fitting curve, the static points that have not been placed in the grid area are placed in the railing area.

[0029] Further, as a preferred technical solution, the process of identifying the static points that can be placed in the railing area includes:

[0030] According to the railing fitting curve, calculate the coordinate positions of the static points that have not been placed in the grid area;

[0031] Compare the calculated coordinate positions with the measured coordinate positions. When the difference between the calculated coordinate positions and the measured coordinate positions is within the preset range, it is determined that the static point is a point on the railing.

[0032] Further, as a preferred technical solution, the static point is a target point detected by the BSD radar with a ground speed of 0, or a target point whose absolute value of the ground speed is less than or equal to the dynamic and static state discrimination threshold.

[0033] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0034] In the present invention, the railing area is divided into multiple grid areas at fixed intervals, and multiple frames of static points detected by the BSD radar are respectively stored in multiple grid areas. The stored data in the grid areas is screened to reduce the amount of computation and storage. At the same time, the grid area is updated in real time or maintained and predicted through the detected static points, and the railing scene division is performed on the static points that have not been added to the grid area, reducing the formation of false targets, and solving the problems of the existing railing recognition scheme that due to the limited number of target points in a single frame and affected by the detection accuracy of the BSD radar, it cannot well distinguish the actual scene of the railing, and the memory consumption and time consumption caused by the large storage amount of multiple frames of target points. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flowchart of the steps of the present invention.

[0036] Figure 2 It is a schematic diagram of the grid area in the vehicle body coordinate system of the present invention;

[0037] Figure 3 It is a schematic diagram of the specific step flow of step S20 of the present invention.

[0038] Figure 4 This is a schematic diagram of the specific steps of step S30 of the present invention.

[0039] Figure 5 This is a schematic diagram of the specific steps of step S40 of the present invention.

[0040] The accompanying drawings are only for illustrative purposes and should not be construed as a limitation of this patent; for better illustration of this embodiment, some components in the accompanying drawings may be omitted, enlarged or reduced, which does not represent the size of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the accompanying drawings may be omitted; the same or similar reference numerals correspond to the same or similar components; the terms describing the positional relationship in the accompanying drawings are only for illustrative purposes and should not be construed as a limitation of this patent. Detailed implementation manners

[0041] The following will elaborate on the preferred embodiments of the present invention in conjunction with the accompanying drawings, so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making the protection scope of the present invention more clearly defined.

[0042] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", "top", "bottom", "inner", "outer", etc. indicating the orientation or positional relationship, they are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, so the terms describing the positional relationship in the accompanying drawings are only for illustrative purposes and should not be construed as a limitation of this patent.

[0043] In addition, if there are terms such as "first", "second", etc., they are only for descriptive purposes, mainly used to distinguish different devices, elements or components (the specific types and structures may be the same or different), and are not used to indicate or imply the relative importance and quantity of the indicated devices, elements or components, and should not be construed as indicating or implying relative importance.

[0044] Embodiment 1

[0045] This embodiment provides a grid-based railing detection method based on a BSD radar to solve the problems of large errors, low efficiency, and large computational complexity in the existing technology of identifying railings by a BSD millimeter-wave radar.

[0046] A grid-based railing detection method based on a BSD radar disclosed in this embodiment, as Figure 1 shown, includes the steps:

[0047] S10. Establish a vehicle body coordinate system, and divide the railing area into multiple grid areas at a fixed interval in the vehicle body coordinate system.

[0048] As Figure 2 shown, taking the single-side radar as an example, this step is specifically: establish a vehicle body coordinate system in the XY direction. In this vehicle body coordinate system, divide the railing area into multiple grid areas at a fixed interval.

[0049] Among them, the size of the grid area is the fixed interval, and the number of grid areas is determined according to the size of the railing area and the size of the grid area.

[0050] S20. Screen the stationary points detected by the BSD radar in each frame, and store them in the divided grid areas, so that each grid area retains at most one stationary point.

[0051] In this step, the stationary point is the target point with a ground speed of 0 detected by the BSD radar.

[0052] In some embodiments, the stationary point can also be the target point whose absolute value of the ground speed detected by the BSD radar is less than or equal to the dynamic and static state discrimination threshold.

[0053] This step is as Figure 3 shown, and specifically includes:

[0054] S201. Screen the stationary points detected by the BSD radar in each frame, and store the stationary points in the divided grid areas according to their coordinate positions respectively.

[0055] S202. Perform edge extraction on the multiple stationary points placed in each grid area, so that each grid area retains at most one stationary point.

[0056] In this step, edge extraction needs to be performed on the stationary points stored in the grid area in each frame to ensure that at most one stationary point is stored in each grid area.

[0057] S30. Perform motion compensation on the stationary points in each grid area according to the vehicle speed of the host vehicle, and update the grid area.

[0058] As Figure 4 shown, this step specifically includes:

[0059] S301. According to the relative position coordinates of the stationary points in each grid area in the vehicle body coordinate system, the vehicle speed of the host vehicle, and the time interval between two adjacent frames, perform position compensation on the stationary points in each grid area, update the coordinate positions of the stationary points, and thus update the grid area.

[0060] S302. According to the distribution of static points in the updated grid areas, perform edge extraction on the grid areas with multiple static points so that each grid area retains at most one static point, and update the grid areas.

[0061] S40. Perform third-order fitting on the static points in all grid areas within a period of time to obtain a railing fitting curve, and determine whether the railing is effective according to the railing fitting curve.

[0062] In this step, polynomial fitting is used for the third-order fitting. There is no fixed limit for the period of time, as long as within a period of time, the target points continuously collected by the BSD radar in multiple frames can be used to determine the effectiveness of the railing.

[0063] This step is as Figure 5 shown, and specifically includes:

[0064] S401. According to the coordinate positions of the static points in the grid areas within a period of time, use polynomial fitting to obtain the railing fitting coefficients.

[0065] In this step, according to the coordinate positions of the static points in all grid areas, the following polynomial is used for fitting to obtain the railing fitting coefficients:

[0066] y = a0 + a1 * x + a2x 2 + a3 * x 3

[0067] where (x, y) represents the coordinate positions of the static points in all grid areas, that is, the coordinate positions of the static points in the first area to the Nth area, and (a0, a1, a2, a3) represents the railing fitting coefficients.

[0068] S402. Obtain the railing fitting curve and the railing fitting deviation according to the railing fitting coefficients.

[0069] The specific implementation process of this step is as follows:

[0070] According to the fitting coefficients obtained in the above steps, referring to the coordinate positions of the static points in all grid areas, the railing fitting curve is obtained. Further, the railing fitting deviation is calculated using the above polynomial.

[0071] In this embodiment, the calculation of the railing fitting deviation is specifically as follows: Calculate the coordinate positions of the static points according to the railing fitting coefficients, and calculate the railing fitting deviation according to the calculated coordinate positions and the measured coordinate positions.

[0072] In this embodiment, an example is given as follows:

[0073] Since the coordinate of the X-axis is a determined value, therefore, the calculation of the railing fitting deviation is as follows:

[0074] According to the railing fitting coefficient and the x-axis coordinate value of the stationary point, calculate the y-axis coordinate value of the stationary point through polynomial, and perform a difference operation between the calculated y-axis coordinate value of the stationary point and the measured y-axis coordinate value of the stationary point, so as to obtain the railing fitting deviation.

[0075] S403. When the railing fitting deviation is less than the preset fitting deviation value, it is determined that the railing fitting is effective.

[0076] In this step, the preset fitting deviation value is the maximum value allowed for the railing fitting deviation set according to the BSD radar road test data.

[0077] Therefore, when the railing fitting deviation calculated in the above steps is within the maximum allowable range of the set railing fitting deviation, it is determined that the railing fitting is effective, and a successfully fitted railing is obtained.

[0078] S50. If the railing is a valid railing, label the stationary points that have not been placed in the grid area.

[0079] This step specifically includes:

[0080] If the railing is a valid railing, according to the railing fitting curve, place the stationary points that have not been placed in the grid area into the railing area.

[0081] In this step, the process of identifying the stationary points that can be placed in the railing area includes:

[0082] S501. According to the railing fitting curve, calculate the coordinate position of the stationary point that has not been placed in the grid area.

[0083] Specifically: Referring to the above steps, extract the railing fitting coefficient and the x-axis coordinate value of the stationary point that has not been placed in the grid area according to the railing fitting curve, and calculate the y-axis coordinate value of the stationary point through polynomial.

[0084] S502. Compare the calculated coordinate position with the measured coordinate position. When the difference between the calculated coordinate position and the measured coordinate position is within the preset range, it is determined that the stationary point is a point on the railing.

[0085] Perform a difference operation between the calculated y-axis coordinate value of the stationary point and the measured y-axis coordinate value of the stationary point that has not been placed in the grid area to obtain the measurement error. When the measurement error is within the preset range, it is determined that the stationary point is a point on the railing.

[0086] Then, divide the stationary points identified through the above steps into the stationary points in the railing area and place them into the corresponding grid areas in the railing area.

[0087] In this embodiment, the stationary points in the grid area do not participate in the tracking of the track. Therefore, the formation of false targets can be reduced, thereby reducing the amount of calculation.

[0088] S60. Update the target list of the BSD radar according to the stationary points in the grid area.

[0089] Since only one-sided radar is taken as an example in this embodiment, therefore, the target lists of the two-sided radars are updated respectively through the above steps.

[0090] Embodiment 2

[0091] This embodiment further discloses a specific implementation manner of the BSD radar for screening and storing stationary points on the basis of Embodiment 1.

[0092] In this embodiment, an example is given for the implementation process of step S20 in Embodiment 1.

[0093] Step S20 includes:

[0094] S201. Screen the stationary points detected by the BSD radar in each frame, and store the stationary points in the divided grid areas according to their coordinate positions.

[0095] S202. Perform edge extraction on the multiple stationary points placed in each grid area, so that at most one stationary point is retained in each grid area.

[0096] The specific implementation process of step S20 is as follows:

[0097] Screen the stationary points from the target points detected by the BSD radar in the previous frame, and store the stationary points in the divided grid areas according to their coordinate positions. If there are multiple stationary points in a grid area, then perform edge extraction on the multiple stationary points in the grid to ensure that only one stationary point is retained in the grid area.

[0098] Repeat the above process: screen the stationary points from the target points detected by the BSD radar in the current frame, and store the stationary points in the divided grid areas according to their coordinate positions. If there are multiple stationary points in a grid area, then perform edge extraction on the multiple stationary points in the grid to ensure that only one stationary point is retained in the grid area.

[0099] In this step, the stationary points screened by the BSD radar in each frame need to be stored. At the same time, after storing the stationary points in each frame, edge extraction is performed on the grid areas where multiple stationary points appear to ensure that only one stationary point is retained in each grid area, so as to reduce the storage amount and subsequent calculation amount of the BSD radar.

[0100] Embodiment 3

[0101] This embodiment further discloses a specific implementation method for the BSD radar to perform motion compensation on the stationary points in the grid area on the basis of Embodiment 1.

[0102] In this embodiment, an example of the implementation process of step S30 in Embodiment 1 is given.

[0103] Step S30 includes:

[0104] S301. Based on the relative position coordinates of the stationary points in each grid area in the vehicle body coordinate system, the vehicle speed of the host vehicle, and the time interval between two adjacent frames, perform position compensation on the stationary points in each grid area, update the coordinate positions of the stationary points, so as to update the grid area.

[0105] In this step, referring to Figure 2 the distribution of the grid areas and the distribution of the stationary points, the following formula is used to perform motion compensation on the stationary points in each grid area:

[0106] X① = |X① + V * deltT|;

[0107] Y① = |Y① + V * deltT|;

[0108] X② = |X② + V * deltT|;

[0109] Y② = |Y② + V * deltT|;

[0110] Among them, (X①, Y①) represents the relative position coordinates of the stationary point ① in the first grid area in the vehicle body coordinate system. Similarly, (X②, Y②) represents the relative position coordinates of the stationary point ② in the second grid area in the vehicle body coordinate system. V represents the vehicle speed of the host vehicle, and deltT represents the time interval between two adjacent frames.

[0111] S302. Based on the distribution of the stationary points in the updated grid area, perform edge extraction on the grid areas with multiple stationary points, so that each grid area retains at most one stationary point, and update the grid area.

[0112] In this step, since the coordinate positions of the compensated stationary points will deviate, therefore, the compensated stationary points may still be in the current grid area, or may have entered other grid areas.

[0113] If the compensated stationary points enter other grid areas, based on the distribution of the stationary points in the compensated grid area, perform edge extraction on the grid areas with multiple stationary points again, so that each grid area retains at most one stationary point, update the grid area, and store the stationary points.

[0114] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, rather than limitations on the implementation manners of the present invention. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the claims of the present invention.

Claims

1. A grid-based railing detection method based on a BSD radar, characterized in that, Including the steps: S10. Establish a vehicle body coordinate system, and divide the railing area into multiple grid areas at a fixed interval in the vehicle body coordinate system; S20. Screen the stationary points detected by the BSD radar for each frame, and store them in the divided grid areas, so that each grid area retains at most one stationary point; S30. Perform motion compensation on the stationary points in each grid area according to the vehicle speed of the host vehicle, and update the grid area; S40. Perform third-order fitting on the stationary points in all grid areas for a period of time to obtain a railing fitting curve, and judge whether the railing is valid according to the railing fitting curve; S50. If the railing is a valid railing, label the stationary points that have not been placed in the grid area; Among them, step S30 specifically includes: S301. According to the relative position coordinates of the stationary points in each grid area in the vehicle body coordinate system, the vehicle speed of the host vehicle, and the interval time between two adjacent frames, perform position compensation on the stationary points in each grid area, update the coordinate positions of the stationary points, so as to update the grid area; S302. According to the distribution of the stationary points in the updated grid area, perform edge extraction on the grid areas with multiple stationary points, so that each grid area retains at most one stationary point, and update the grid area.

2. The grid railing detection method based on a BSD radar according to claim 1, characterized in that, Step S20 specifically includes: S201. Screen the stationary points detected by the BSD radar for each frame, and store the stationary points in the divided grid areas according to their coordinate positions respectively; S202. Perform edge extraction on the multiple stationary points placed in each grid area, so that each grid area retains at most one stationary point.

3. The grid railing detection method based on a BSD radar according to claim 2, characterized in that, In step S202, edge extraction needs to be performed on the stationary points stored in the grid area for each frame to ensure that at most one stationary point is stored in each grid area.

4. A grid-based railing detection method based on a BSD radar according to claim 1, characterized in that Step S40 specifically includes: S401. According to the coordinate positions of the stationary points in the grid areas for a period of time, use polynomial fitting to obtain railing fitting coefficients; S402. Obtain a railing fitting curve and a railing fitting deviation according to the railing fitting coefficients; S403. When the railing fitting deviation is less than the preset fitting deviation value, it is determined that the railing fitting is valid.

5. A grid-based railing detection method based on a BSD radar according to claim 4, characterized in that, The preset fitting deviation value is the maximum value allowed for the railing fitting deviation set according to the BSD radar road test data.

6. The grid railing detection method based on a BSD radar according to claim 1, characterized in that, Step S50 specifically includes: If the railing is a valid railing, according to the railing fitting curve, put the stationary points that have not been placed in the grid area into the railing area.

7. A grid-based railing detection method based on a BSD radar according to claim 6, characterized in that, The identification process of the stationary points that can be placed in the railing area includes: According to the railing fitting curve, calculate the coordinate positions of the stationary points that have not been placed in the grid area; Compare the calculated coordinate positions with the measured coordinate positions. When the difference between the calculated coordinate positions and the measured coordinate positions is within the preset range, it is determined that the stationary point is a point on the railing.

8. A grid-based railing detection method based on a BSD radar according to claim 6, characterized in that, The stationary point is a target point detected by the BSD radar with a ground speed of 0, or a target point whose absolute value of the ground speed is less than or equal to the dynamic and static state discrimination threshold.

Images