CN114399460A - Method and system for detecting depth of accumulated water on road surface and vehicle - Google Patents
Method and system for detecting depth of accumulated water on road surface and vehicle Download PDFInfo
- Publication number
- CN114399460A CN114399460A CN202111456629.3A CN202111456629A CN114399460A CN 114399460 A CN114399460 A CN 114399460A CN 202111456629 A CN202111456629 A CN 202111456629A CN 114399460 A CN114399460 A CN 114399460A
- Authority
- CN
- China
- Prior art keywords
- water
- depth
- ponding
- point
- vehicle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
Abstract
The invention discloses a method for detecting depth of surface gathered water, which comprises the following steps: 100: collecting image information of a water accumulation road surface; 200: processing the image information to obtain pixel coordinate information of accumulated water edge points; 300: acquiring relative position coordinates of each accumulated water marginal point and a vehicle; 400: obtaining world coordinates of the ponding edge points in the high-precision map based on the relative position coordinates and the position coordinates of the vehicle in the high-precision map, and further obtaining each point of an area in a closed curve formed by the ponding edge points and the elevation information of the ponding edge points; 500: obtaining depth distribution information of each point in the ponding area according to the altitude coordinate difference between each ponding edge point and each point in the area in the closed curve; 600: drawing a water depth moire pattern of the ponding area based on the depth distribution information; 700: and displaying or projecting the water depth moire pattern on a display terminal. Correspondingly, the invention also discloses a surface water depth detection system for implementing the method, which comprises an image acquisition device, a processing module and a display terminal.
Description
Technical Field
The invention relates to an intelligent detection method, in particular to a depth detection method for surface gathered water.
Background
In recent years, smart cars with an automatic driving function have come into the field of view of the public. In such an intelligent vehicle, various sensors are provided, which can sense the surrounding environment by means of the sensors, and can effectively assist driving.
When extreme weather, such as heavy rainfall, etc., the low-lying road sections around the road, especially on the road, may be submerged by water; when the vehicle is trapped in water, the anchorage is easily caused. In addition, when nearby rivers and lakes are covered, water accumulation pavements are easy to appear; when the depth of the accumulated water is large, the accumulated water possibly enters an exhaust pipe of the engine due to the water flowing of the automobile, and the engine is flamed out; if water enters the cylinder, it can also cause mechanical damage to the engine valve train and the cylinder head.
For an intelligent vehicle with an automatic driving function, an Advanced Driver Assistance System (ADAS) plays an important role, which requires an advanced route planning according to a travelable area, and even in the case of an assisted driving, a driver knows a surface water depth distribution map in advance, and can select to wade through a shallow area according to information of the surface water depth distribution map.
Therefore, it is important to extract the water depth distribution information and to know which areas have the water depths of the road surface as drivable areas and which areas have the water depths of the road surface as undrivable areas.
Currently, some researchers have made relevant studies to this problem and have obtained certain research results, but the effect of the existing technical solutions in practical application is not ideal:
for example: chinese patent publication No. CN109741391A, published as 5/10/2019 entitled "method, apparatus, and storage medium for detecting depth of water accumulated on road surface". According to the technical scheme, a camera is used for shooting pictures of the side face of the vehicle, wheel edge information is obtained from the images, and the depth of accumulated water is obtained. This technical scheme can judge the water level degree of depth according to tire draft position, but it can not judge the place ahead ponding degree of depth in advance before the vehicle tire wades for the vehicle drives into the deep water district of advance and retreat dilemma easily under the condition of knowing, and then causes the breakdown.
For another example: publication No. CN113168535A, publication date 2021, 7 months 23 days; the name of the Chinese patent document is 'a method and a device for determining the depth of accumulated water'. The technical scheme includes that road information of the maximum depth is determined by acquiring first edge information of a ponding road and combining road slope angle information; the method sets the ponding area as an ideal ramp and calculates the maximum ponding depth according to the trigonometric function relationship. The method includes determining the maximum water depth position by adopting gradient change, and calculating the maximum water depth according to the maximum water depth position. However, in this solution, especially the logic of trigonometric function relation is based on the assumption of ideal plane ramp road surface, and the maximum water depth position is calculated. In fact, when an irregular or uneven road surface is encountered, even if the road surface is small, the vehicle can shake and bump, and the assumption of an ideal plane of the road surface can be broken, so that a distance error can be caused. In addition, the road of the ponding area is not completely an ideal plane, for example, the situation of the road non-ideal plane occurs, which causes irregular altitude distribution, and further causes irregular depth distribution of the ponding. Therefore, the water depth of only one maximum point cannot meet the requirements of users; under certain conditions, even if the maximum water depth of a part of road area is larger than the warning depth, certain bypassing trafficability still exists.
Therefore, if the depth distribution of the ponding area can be acquired, and a very intuitive water depth moire pattern method is used for displaying or projecting the water depth moire pattern on a vehicle display screen or a front windshield of a vehicle, convenience can be provided for a driver of the vehicle; meanwhile, under the condition of obtaining detailed water accumulation depth distribution, the method is more favorable for drawing a travelable area by an automatic driving vehicle algorithm, so that the automatic driving capability of the vehicle is improved.
Therefore, the invention is expected to obtain the method for detecting the depth of the surface gathered water, which can effectively calculate the depth distribution of the surface gathered water, draw the water depth moire pattern of the water accumulation area according to the depth distribution information, and further display or project the water depth moire pattern on the display terminal.
Disclosure of Invention
One of the purposes of the invention is to provide a method for detecting depth of surface water, which can effectively calculate depth distribution of surface water, draw a water depth moire pattern of a water accumulation area according to depth distribution information, and further display or project the water depth moire pattern on a display terminal.
In order to achieve the purpose, the invention provides a depth detection method of surface gathered water, which comprises the following steps:
100: collecting image information of a water accumulation road surface;
200: processing the image information to obtain pixel coordinate information of the edge point of the accumulated water;
300: acquiring relative position coordinates of each accumulated water marginal point and a vehicle;
400: obtaining world coordinates of the ponding edge points in the high-precision map based on the relative position coordinates and the position coordinates of the vehicle in the high-precision map, and further obtaining elevation information of each point of an area in a closed curve formed by the ponding edge points and the elevation information of the ponding edge points;
500: acquiring depth distribution information of each point in the ponding area based on the altitude coordinate difference between each ponding edge point and each point in the area in the closed curve;
600: drawing a water depth moire pattern of the ponding area based on the depth distribution information;
700: and displaying or projecting the water depth moire pattern on a display terminal.
In the technical scheme of the invention, the inventor creatively designs a method for detecting the depth of the accumulated water on the road surface, which effectively solves the problem of acquiring the drivable area of the accumulated water area before a vehicle enters the accumulated water area. By adopting the method for detecting the depth of the surface gathered water, the depth distribution of the surface gathered water can be effectively calculated, the water depth moire pattern of the water gathering area is drawn according to the depth distribution information, and the water depth moire pattern is displayed or projected on a display terminal (such as a vehicle-mounted display screen or a front windshield of a vehicle).
In the technical scheme of the invention, the depth detection method for the surface gathered water can avoid risks caused by terminal wading, and the water depth moire pattern can visually help a driver to obtain information of a drivable area of the road. The method can also be combined with roadbed equipment and applied to the car networking, such as the vehicle external connection V2X, so that the method further assists a driver and provides early warning information.
Meanwhile, the method for detecting the depth of the water accumulated on the road surface can improve the capability of distinguishing the travelable area by the automatic driving automobile, and is more favorable for drawing the travelable area by an automatic driving automobile algorithm under the condition of obtaining the detailed water accumulated depth distribution, so that the automatic driving or auxiliary driving capability of the automobile is improved.
In some embodiments, the method for detecting depth of surface water may also be used in combination with VR technology and head-up display technology to project a water depth moire pattern onto a front windshield of a vehicle.
Further, in the method for detecting depth of surface water accumulation according to the present invention, in step 200, pixel coordinate information of an edge point of water accumulation is extracted from the image information by using a semantic segmentation or instance segmentation method.
Further, in the method for detecting depth of standing water in a road surface according to the present invention, the step 200 further includes: it is determined whether or not it is a stagnant area based on the image information, and if it is determined as "yes", step 300 is performed.
Further, in the method for detecting depth of accumulated water on the road surface, a convolutional neural network classification model is adopted to judge whether the area is an accumulated water area.
Further, in the method for detecting depth of water accumulated on a road surface according to the present invention, in step 300, a visual depth estimation algorithm is used to obtain the relative position coordinates of each water accumulated edge point and the vehicle.
Further, in the depth detection method of the surface water according to the present invention, in step 400, the elevation information is directly obtained from a high-precision map; or the altitude information is obtained by calculation based on the slope angle information in the high-precision map.
Further, in the method for detecting depth of accumulated water on a road surface according to the present invention, in step 600, color interpolation is performed on each point in the accumulated water area by using a pixel interpolation method to draw the water depth moire pattern.
Further, in the method for detecting depth of surface water according to the present invention, in step 700, the water depth moire pattern is displayed on a vehicle display screen of the vehicle, or projected on a front windshield of the vehicle by using a VR technology.
Accordingly, another object of the present invention is to provide a depth detection system for road surface water, which can be used to implement the above-mentioned depth detection method for road surface water.
In order to achieve the above object, the present invention provides a depth detection system for surface water, which can be used for implementing the above depth detection method for surface water of the present invention, wherein the depth detection system for surface water comprises an image acquisition device, a processing module and a display terminal; wherein the image capturing device performs step 100, and the processing module performs step 200 and step 700.
In addition, the invention also aims to provide a vehicle capable of detecting the depth of the surface water, and in order to achieve the aim, the invention provides a vehicle which is provided with the surface water depth detection system.
Compared with the prior art, the method, the system and the vehicle for detecting the depth of the surface gathered water have the advantages and beneficial effects as follows:
different from the prior art, the inventor creatively designs a method for detecting the depth of the accumulated water on the road surface, which effectively solves the problem of acquiring the drivable area of the accumulated water area before a vehicle enters the accumulated water area. By adopting the method for detecting the depth of the surface gathered water, the depth distribution of the surface gathered water can be effectively calculated, the water depth moire pattern of the water gathering area is drawn according to the depth distribution information, and the water depth moire pattern is displayed or projected on a display terminal (such as a vehicle-mounted display screen or a front windshield of a vehicle).
The method for detecting the depth of the surface gathered water can avoid risks caused by terminal wading, and the water depth moire pattern can visually help a driver to obtain information of a drivable area of the road. Meanwhile, the method is beneficial to drawing a travelable area by an automatic driving vehicle algorithm, and the automatic driving or auxiliary driving capacity of the vehicle is improved.
In some embodiments, the method for detecting depth of water accumulated on a road surface according to the present invention may further be applied to a vehicle networking, such as an external vehicle connection V2X, in combination with a roadbed device, so as to further assist a driver and provide warning information.
Correspondingly, the surface water depth detection system for implementing the surface water depth detection method and the surface water depth detection system have the advantages and the beneficial effects.
Drawings
Fig. 1 schematically shows a top view of an application scenario of the depth detection system for road surface water according to an embodiment of the present invention.
Fig. 2 schematically shows a flow chart of the steps of the method for detecting depth of surface water according to one embodiment of the present invention.
Fig. 3 schematically shows a flowchart of specific steps of extracting pixel coordinate information of a ponding edge point and determining whether a ponding area exists in step 200 according to the depth detection method for the ponding in the road surface of the invention.
Fig. 4 schematically shows a pixel matrix schematic diagram output by an example segmentation algorithm model in an embodiment of the method for detecting depth of surface water according to the present invention.
FIG. 5 schematically shows a method for detecting depth of water accumulated on a road surface according to the present invention, in one embodiment, relative position coordinates dx of each water accumulation edge point and a vehicle are obtained through a visual depth estimation algorithmiAnd dyiIs described.
Fig. 6 schematically shows the calculation of elevation information for each point of the area within the closed curve formed by the edge points of the water based on the slope angle information.
FIG. 7 schematically illustrates a method for roughly screening out high-precision map points in the waterlogged area.
FIG. 7 schematically illustrates a method for roughly screening out high-precision map points in the waterlogged area.
FIG. 8 schematically illustrates a method of fine screening of high-precision map points within an area of stagnant water.
Fig. 9 schematically shows a schematic diagram of depth distribution of each point in a calculated ponding area according to the method for detecting depth of ponding water in a road surface of the invention.
Detailed Description
The method, system and vehicle for detecting depth of water in a road surface according to the present invention will be further explained and explained with reference to the drawings and the specific embodiments, however, the explanation and explanation should not be construed as an inappropriate limitation to the technical solution of the present invention.
In the present invention, the surface water depth detection system of the present invention may include: image acquisition device, processing module and display terminal. The image acquisition device can acquire image information of the ponding road surface; the processing module can process the image information to draw and obtain a water depth moire pattern of the ponding area, and display or project the water depth moire pattern on the display terminal.
It should be noted that the system for detecting depth of surface water can not only be applied to vehicles, but also be applied to devices with cameras, such as transportation devices (e.g., transportation cameras), unmanned aerial vehicles, railcars, water depth measuring devices, handheld devices (e.g., mobile phones) and the like, in some other embodiments.
Fig. 1 schematically shows a top view of an application scenario of the depth detection system for road surface water according to an embodiment of the present invention.
As shown in fig. 1, in this embodiment, a surface water depth detection system may be installed on a vehicle, an image acquisition device may be selected as a camera, the vehicle may acquire image information of a road water accumulation region in real time through the camera, and transmit the image information to a processing module, the processing module may process and calculate water depth distribution, draw a water depth moire pattern (equal-depth line), and display or project the water depth moire pattern on a display terminal (for example, a vehicle-mounted display screen or a front windshield of the vehicle).
Correspondingly, an Advanced Driver Assistance System (ADAS) of the vehicle can determine a danger zone, a warning zone and a travelable zone of the surface water according to the depth distribution of the water accumulation area and the water depth moire pattern and by combining the vehicle chassis height information so as to plan the detour route of the vehicle later.
In the present embodiment, the road surface water depth detection system according to the present invention implements the road surface water depth detection method shown in fig. 2 described below, thereby achieving the above-described effects.
Fig. 2 schematically shows a flow chart of the steps of the method for detecting depth of surface water according to one embodiment of the present invention.
As shown in fig. 2, in the present embodiment, the method for detecting depth of standing water according to the present invention includes the steps of:
step 100: and collecting image information of the ponding road surface through the camera.
Step 200: and processing the image information to obtain pixel coordinate information of accumulated water edge points, and automatically judging whether the accumulated water area exists or not by combining a deep learning image binary model.
In step 200 of the present invention, pixel coordinates of edge points of the water can be extracted based on the image information, and are expressed as { (w)1,h1),(w2,h2)，(w3,h3),(w4,h4) … …, and determining whether a water accumulation zone existsThe specific step flow can be seen in fig. 3.
Fig. 3 schematically shows a flowchart of specific steps of extracting pixel coordinate information of a ponding edge point and determining whether a ponding area exists in step 200 according to the depth detection method for the ponding in the road surface of the invention.
As shown in fig. 3, in this embodiment, the step 200 according to the present invention may further include steps 201 to 209 as follows:
step 201: the image information is subjected to resize processing in Open CV (the Open Source Computer Vision library is an Open-Source-based cross-platform Computer Vision library which can run on Linux, Windows and Mac OS operating systems, is light and efficient, consists of a series of C functions and a small number of C + + classes, provides interfaces of languages such as Python, Ruby, MATLAB and the like, and realizes image processing and a plurality of general algorithms in the aspect of Computer Vision). For example, a 256 × 256-size image can be obtained based on the image information resize.
Step 202: and sending the image after resize into a trained example segmentation deep learning model.
Step 203: the model output results in a binarized pixel matrix with pixels 0 and 1. Referring to fig. 4 for details, fig. 4 schematically shows a pixel matrix schematic diagram output by an example segmentation algorithm model in an embodiment of the method for detecting depth of surface water according to the present invention. As shown in fig. 4, the pixel of the water accumulation region is "1", and the pixel of the non-water accumulation region is "0".
Step 204: and extracting the pixel semantic edge point position through a specific algorithm to obtain the pixel coordinate information of the ponding edge point.
For example, in one embodiment, the pixel coordinates of the first non-zero pixel in the pixel matrix may be extracted in row and column and in forward and reverse order, respectively. Similar edge extraction methods are many, and for example, edge information can be extracted by a Canny edge detection algorithm. In the present invention, the method of semantic edge extraction of pixels is not specified.
In the present inventionIn step 204, the accumulated water edge points { n } can be obtained finally1,n2,n3,n4… … { (w)1,h1),(w2,h2)，(w3,h3),(w4,h4),……}。
Step 205: maximum value max w according to pixel edgei},max{hiAnd minimum min wi},min{hiAnd further generating a cutting frame range of the region of interest in the original picture, and recording the cutting frame range as box { w, h }.
Step 206: the image cropped in step 205 and multiplied by h is processed by a resize module in open CV, for example, the resize is a 128 × 128 size image.
Step 207: and (4) sending the image subjected to resize in the step 206 into a deep learning image classification model in a convolutional neural network classification model to obtain a judgment result.
Step 208: if the judgment result is yes, namely a water accumulation area exists, starting a subsequent water accumulation depth distribution calculation flow immediately, and performing the subsequent step 300; otherwise, the process is terminated and the subsequent step 300 is not performed.
Step 300: and obtaining the relative position coordinates of each ponding marginal point and the vehicle through a visual depth estimation algorithm.
In the above step 300 of the present invention, the relative coordinates (dx) of the vehicle can be calculated from the pixel coordinates of each edge point by the visual depth estimation algorithmi,dyi) The specific process can be seen in table 5.
FIG. 5 schematically shows a method for detecting depth of water accumulated on a road surface according to the present invention, in one embodiment, relative position coordinates dx of each water accumulation edge point and a vehicle are obtained through a visual depth estimation algorithmiAnd dyiIs described.
As shown in fig. 5, in this embodiment, step 300 may specifically include the following steps:
step 301: images of left and right cameras of the waterlogged area are obtained through the double-sided camera.
Step 302: through a homography matrix between the phasing plane of the left camera and the phasing plane of the right camera.
Step 303: and carrying out homography transformation on the images of the left camera and the right camera by utilizing the homography matrix, and determining the displacement offset of each pixel point in the images by the homography transformation.
Step 304: and matching pixel points of the transformed image and another image through a stereo matching algorithm, and calculating the parallax between the pixel points which are successfully matched.
Step 305: the parallax calculated in step 304 is compensated by the displacement shift amount in step 303.
Step 306: the actual disparity between the left and right images is confirmed.
Step 307: according to the actual parallax, combining a trigonometric function, calculating depth information of the accumulated water edge pixel points; and decomposing and calculating the relative position coordinate dx of the accumulated water edge pixel point relative to the self-vehicle under the self-vehicle coordinate systemiAnd dyi。
It should be noted that there are many methods for the visual depth estimation algorithm, and the present invention does not limit the method. In the present embodiment, the distance between each ponding edge point and the camera of the vehicle is calculated from the parallax of the left and right cameras using binocular stereoscopic depth estimation. The input of this step is the edge point pixel coordinates { (w) obtained in step 2001,h1),(w2,h2)，(w3,h3),(w4,h4) … … }; obtaining the relative position coordinates of each ponding marginal point and the vehicle through a visual depth estimation algorithm, and recording as { (dx)1,dy1),(dx2,dy2)，(dx3,dy3),(dx4,dy4),……}。
Step 400: and acquiring world coordinates of the ponding edge points in the high-precision map based on the relative position coordinates of the ponding edge points and the vehicle and the position coordinates of the vehicle in the high-precision map, and further acquiring altitude information of each point of an area in a closed curve formed by the ponding edge points and the altitude information of the ponding edge points.
In the embodiment, the position coordinates of the vehicle in the high-precision map can be obtained based on GPS positioning, and the world coordinates of the edge points of the ponding in the high-precision map can be calculated by combining the GPS positioning information of the vehicle in the high-precision map; of course, in some other embodiments, the position coordinates of the vehicle in the high-precision map can also be obtained through Beidou positioning or other common positioning modes.
In the above step 400 of the present invention, the position coordinates of the vehicle in the high-precision map are superimposed on the relative position coordinates of each ponding edge point and the vehicle obtained in step 300, so as to obtain ponding edge points { n }1,n2,n3,n4… … world coordinates in high-precision maps, world coordinates (x) of edge points of each water accumulationi,yi) Can be respectively described as { (x)1,y1),(x2,y2)，(x3,y3),(x4,y4),……}。
In the invention, the world coordinate (x) of each ponding edge point is based oni,yi) The (x) of each ponding edge point can be further calculatedi,yi) Altitude coordinate zi. In some embodiments, a search radius R may be set, and the average altitude of the coordinates of points on the high-precision map within the radius R may be regarded as the edge point (x) of the ponding wateri,yi) Altitude coordinate zi。
Correspondingly, according to the obtained world coordinates of each ponding edge point in the high-precision map { (x)1,y1),(x2,y2)，(x3,y3),(x4,y4) … …, each high-precision map Point { Point ] of the area in the closed curve formed by the edge points of the accumulated water can be screened out through a specific screening algorithm1,Point2,Point3,Point4… …, and then obtaining the altitude coordinates of each point in the closed curve formed by the edge points of the ponding water (namely, in the ponding area), and marking as { z }1z2,z3,z4,……}。
It should be noted that if high precision is requiredThe elevation information is not included in the map, and the slope angle information theta is providediThen, the elevation of the edge point of the accumulated water can be set to be 0, and the relative position delta l of each point is usediSum slope angle thetaiCalculating the relative altitude delta z of each point relative to the reference altitude by utilizing trigonometric function relationiThe specific formula is Deltazi＝Δli*sinθi. Detailed description of the drawings referring to fig. 6, fig. 6 schematically shows calculation of elevation information for each point of a region within a closed curve formed by water edge points based on slope angle information.
It should be noted that, in the above-mentioned technical solution of step 400, there are many methods for screening points in the ponding area, one rough method can be seen in the schematic diagram shown in fig. 7, and fig. 7 schematically shows one rough method for screening high-precision map points in the ponding area.
As shown in fig. 7, the specific way is to filter out a rectangular box directly according to the maximum coordinate and the minimum coordinate of each point. First calculate xmin＝min{xi},ymin＝min{yi}，xmax＝max{xi},ymax＝max{yi}; then screening out the satisfied (x)min<xi<xmax) And (y)min<yi<ymax) High-precision map Point of (Point)1,Point2,Point3,Point4… …, and extracting the altitude coordinates of each high-precision map point in the ponding area, and recording as { z }1,z2,z3,z4,……}。
Of course, on the basis of the screening method shown in fig. 7, more detailed region screening can be performed, and the specific screening step flow can be referred to fig. 8 below.
FIG. 8 schematically illustrates a method of fine screening of high-precision map points within an area of stagnant water.
As shown in fig. 8, for example: firstly, the world coordinates of each ponding edge point in a high-precision map { (x)1,y1),(x2,y2)，(x3,y3),(x4,y4) … … } latitude coordinate y1，y2，y3，y4…; for example, the longitude and the latitude are divided into N intervals, and the length of each interval is Δ ═ y _ max-y _ min)/N, see fig. 8. Then for a certain interval numbered i, its dimension coordinate interval is Δ yi＝{ymin+i*Δ，yminPlus (i +1) × Δ, where i is any natural number from 0 to N-1 onwards. According to Δ yiThe longitude minimum value and the longitude maximum value Deltax of the edge point of the ponding water in the interval can be further obtainedi＝{xmin,xmaxIs further passed by Δ xiAnd Δ yiScreening out each high-precision map Point { Point within a section1,Point2,Point3,Point4… …, and extracting the altitude coordinates of each high-precision map point in the ponding area, and recording as { z }1z2,z3,z4,……}。
It should be noted that the screening methods shown in fig. 7 and 8 are only two examples of the method for screening high-precision map points in the ponding area and extracting elevation information, and are used to illustrate the feasibility of screening out each point of the area in the closed curve formed by the ponding edge points according to the world coordinates of the ponding edge points. In fact, in the present invention, the screening method is not particularly limited.
Step 500: and obtaining the depth distribution information of each point in the ponding area based on the altitude coordinate difference between each ponding edge point and each point in the area in the closed curve.
Fig. 9 schematically shows a schematic diagram of depth distribution of each point in a calculated ponding area according to the method for detecting depth of ponding water in a road surface of the invention.
Referring to fig. 9, in the present embodiment, the step 500 of the present invention may specifically include the following steps 501 and 502:
step 501: based on the obtained altitude coordinate z of each ponding edge pointiCalculating the altitude coordinate of the accumulated water
Step 502: based on the acquired elevation coordinate z of the ponding0Subtracting the altitude coordinate (z) of each point in the closed curve formed by the edge points of the accumulated water1z2,z3,z4… … }, i.e. hi＝z0-ziTo obtain the water depth distribution { h) of each point1h2,h3,h4,……}。
It should be noted that the water depth distribution { h ] calculated by the above algorithm1h2,h3,h4… …, which can be matched in conjunction with the world coordinates of the points of the waterlogged area in a high-precision map. In the invention, through coordinate conversion, the world coordinates of each point in the ponding area in the high-precision map can be converted into pixel coordinates, so that the depth distribution information { h) of each pixel point in the ponding area is obtained1h2,h3,h4,……}。
Step 600: and drawing a water depth moire pattern of the ponding area based on the depth distribution information.
Based on the depth distribution information of each point in the ponding area obtained in the step 500, a water depth moire pattern of the ponding area can be drawn by combining a linear interpolation pixel completion method in openCV, which is shown in an application scene top view shown in fig. 1.
In the present embodiment, the water depth moire map may be drawn by performing color interpolation on each point in the ponding region by using a pixel interpolation method. In particular, the depth h may be established byiAnd a pixel value Ci(real number between 0 and 255) linear mapping relation Ci＝f(hi) The functional relationship may be set as a linear monotonically increasing function, for example, the deeper the water is, the darker the color is, and the mapping relationship between the water depth value and the pixel value in the water depth moire pattern is not limited in the present invention.
In addition, it should be noted that, when drawing the water depth moire map, the high-precision map information may be used as the foundation base plate, and used as the bottom layer pixels of the water depth moire map, and then the pixel value C is superimposed on the map layer according to the above-mentioned pixel interpolation methodiAnd then drawn intoAnd (4) a water depth moire pattern on the basis of a high-precision map top view.
In addition, different from the above technical solution, in some other embodiments, the world coordinates of the water accumulation area points on the high-precision map may be converted into the pixel coordinates of the real-time real scene picture acquired by the front camera through coordinate conversion by using the pixels of the real picture acquired by the front camera as a base board. And drawing a water depth moire pattern according to the pixel interpolation method on the basis of the ponding road image of the camera. The water depth moire pattern obtained by the method is more intuitive.
It should be noted that no matter which of the two methods is adopted to draw the water depth moire pattern, the pixel coordinates of the ponding region points projected onto the picture are sparse, so that when the water depth moire pattern is drawn, a related interpolation algorithm can be further adopted to perform moire color interpolation completion on the pixel values of the adjacent pixel coordinate points. In the present invention, such a conventional interpolation algorithm is not particularly limited.
Step 700: and displaying or projecting the water depth moire pattern on a display terminal.
It should be noted that the display terminal may be a vehicle display screen or a vehicle front windshield, may be on the vehicle display screen of the vehicle, or may be projected onto the vehicle front windshield by combining with the VR technology.
Of course, in some other embodiments, the display terminal may also be a mobile display terminal such as a mobile phone or a tablet computer, and the method for detecting depth of surface water according to the present invention may combine with the internet of things technology to display the water depth moire pattern on the display terminal.
It should be noted that the combination of the features in the present application is not limited to the combination described in the claims of the present application or the combination described in the embodiments, and all the features described in the present application may be freely combined or combined in any manner unless contradicted by each other.
It should also be noted that the above-mentioned embodiments are only specific embodiments of the present invention. It is apparent that the present invention is not limited to the above embodiments and similar changes or modifications can be easily made by those skilled in the art from the disclosure of the present invention and shall fall within the scope of the present invention.
Claims (10)
1. A method for detecting depth of surface gathered water is characterized by comprising the following steps:
100: collecting image information of a water accumulation road surface;
200: processing the image information to obtain pixel coordinate information of the edge point of the accumulated water;
300: acquiring relative position coordinates of each accumulated water marginal point and a vehicle;
400: obtaining world coordinates of the ponding edge points in the high-precision map based on the relative position coordinates and the position coordinates of the vehicle in the high-precision map, and further obtaining elevation information of each point of an area in a closed curve formed by the ponding edge points and the elevation information of the ponding edge points;
500: acquiring depth distribution information of each point in the ponding area based on the altitude coordinate difference between each ponding edge point and each point in the area in the closed curve;
600: drawing a water depth moire pattern of the ponding area based on the depth distribution information;
700: and displaying or projecting the water depth moire pattern on a display terminal.
2. The method for detecting depth of water accumulation in road surface according to claim 1, wherein in step 200, pixel coordinate information of edge points of water accumulation is extracted from the image information by adopting a semantic segmentation method or an example segmentation method.
3. The method for detecting depth of standing water according to claim 1, wherein the step 200 further comprises: it is determined whether or not it is a stagnant area based on the image information, and if it is determined as "yes", step 300 is performed.
4. The method for detecting depth of water accumulation on a road surface according to claim 3, wherein a convolutional neural network classification model is used to determine whether the water accumulation region is present.
5. The method of claim 1, wherein in step 300, a vision depth estimation algorithm is used to obtain the relative position coordinates of each water accumulation edge point and the vehicle.
6. The method for detecting depth of standing water according to claim 1, wherein in step 400, the elevation information is directly obtained from a high-precision map; or the altitude information is obtained by calculation based on the slope angle information in the high-precision map.
7. The method for detecting depth of water accumulation according to claim 1, wherein in step 600, color interpolation is performed on each point in the water accumulation area by using a pixel interpolation method to draw the water depth moire pattern.
8. The method for detecting the depth of the water accumulation according to claim 1, wherein in step 700, the water depth moire pattern is displayed on a vehicle display screen of the vehicle or projected on a front windshield of the vehicle through a VR technology.
9. A surface water depth detection system, which implements the surface water depth detection method according to any one of claims 1 to 8, wherein the surface water depth detection system includes an image acquisition device, a processing module, and a display terminal; wherein the image capturing device performs step 100, and the processing module performs step 200 and step 700.
10. A vehicle characterized by being equipped with the surface water depth detection system as claimed in claim 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111456629.3A CN114399460A (en) | 2021-11-30 | 2021-11-30 | Method and system for detecting depth of accumulated water on road surface and vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111456629.3A CN114399460A (en) | 2021-11-30 | 2021-11-30 | Method and system for detecting depth of accumulated water on road surface and vehicle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114399460A true CN114399460A (en) | 2022-04-26 |
Family
ID=81225705
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111456629.3A Pending CN114399460A (en) | 2021-11-30 | 2021-11-30 | Method and system for detecting depth of accumulated water on road surface and vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114399460A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114758139A (en) * | 2022-06-16 | 2022-07-15 | 成都鹏业软件股份有限公司 | Foundation pit accumulated water detection method |
-
2021
- 2021-11-30 CN CN202111456629.3A patent/CN114399460A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114758139A (en) * | 2022-06-16 | 2022-07-15 | 成都鹏业软件股份有限公司 | Foundation pit accumulated water detection method |
| CN114758139B (en) * | 2022-06-16 | 2022-10-21 | 成都鹏业软件股份有限公司 | Method for detecting accumulated water in foundation pit |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11967109B2 (en) | Vehicle localization using cameras | |
| CN110443225B (en) | Virtual and real lane line identification method and device based on feature pixel statistics | |
| CN112184818B (en) | Vision-based vehicle positioning method and parking lot management system applying same | |
| CN102682292B (en) | Method based on monocular vision for detecting and roughly positioning edge of road | |
| CN107133985B (en) | Automatic calibration method for vehicle-mounted camera based on lane line vanishing point | |
| CN110501018B (en) | Traffic sign information acquisition method for high-precision map production | |
| US20200041284A1 (en) | Map road marking and road quality collecting apparatus and method based on adas system | |
| JP6442834B2 (en) | Road surface height shape estimation method and system | |
| Nedevschi et al. | A sensor for urban driving assistance systems based on dense stereovision | |
| CN106324618B (en) | Realize the method based on laser radar detection lane line system | |
| CN104008377A (en) | Ground traffic sign real-time detection and recognition method based on space-time correlation | |
| CN111209780A (en) | Lane line attribute detection method and device, electronic device and readable storage medium | |
| Liu et al. | Development of a vision-based driver assistance system with lane departure warning and forward collision warning functions | |
| EP2743861B1 (en) | Method and device for detecting continuous object in disparity direction based on disparity map | |
| CN103204104B (en) | Monitored control system and method are driven in a kind of full visual angle of vehicle | |
| CN109635737A (en) | Automobile navigation localization method is assisted based on pavement marker line visual identity | |
| CN114898296A (en) | Bus lane occupation detection method based on millimeter wave radar and vision fusion | |
| CN111694011A (en) | Road edge detection method based on data fusion of camera and three-dimensional laser radar | |
| WO2021017211A1 (en) | Vehicle positioning method and device employing visual sensing, and vehicle-mounted terminal | |
| CN109241855B (en) | Intelligent vehicle travelable area detection method based on stereoscopic vision | |
| Wu et al. | Adjacent lane detection and lateral vehicle distance measurement using vision-based neuro-fuzzy approaches | |
| CN107220632B (en) | Road surface image segmentation method based on normal characteristic | |
| CN114120283A (en) | Method for distinguishing unknown obstacles in road scene three-dimensional semantic segmentation | |
| CN106803073B (en) | Auxiliary driving system and method based on stereoscopic vision target | |
| JP2005217883A (en) | Method for detecting flat road area and obstacle by using stereo image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |