CN108171695A - A kind of express highway pavement detection method based on image procossing - Google Patents
A kind of express highway pavement detection method based on image procossing Download PDFInfo
- Publication number
- CN108171695A CN108171695A CN201711480274.5A CN201711480274A CN108171695A CN 108171695 A CN108171695 A CN 108171695A CN 201711480274 A CN201711480274 A CN 201711480274A CN 108171695 A CN108171695 A CN 108171695A
- Authority
- CN
- China
- Prior art keywords
- edge
- point
- straight line
- line
- image
- 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
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
- 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
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of express highway pavement detection method based on image procossing, including：An at least width video frame images for continuous acquisition video file, and obtain target video frame image according to an at least width video frame images；Edge detection is carried out to target video frame image, obtains the edge image for including road edge pixel point；Edge image is scanned, road area is obtained, and road area laterally divide and obtains subregion, and each sub-regions are detected using probability Hough transformation, obtains road edge line segment；All edge line segments in the subregion on road area top solve vanishing point, and the intermediate control point of each non-bottom end subregion and the boundary point of lowermost end subregion are determined with the presence or absence of intersection point according to the straight line of maximum slope in each sub-regions and the straight line of slope minimum；According to intermediate control point, boundary point, vanishing point, left and right track edge line is drawn.Using the embodiment of the present invention, the adaptability to bend scene is improved.
Description
Technical field
The present invention relates to pavement detection field, more particularly to a kind of express highway pavement detection side based on image procossing
Method.
Background technology
The intelligent video monitoring of highway usually only focuses on road surface region in picture, but day is often included in monitoring image
The backgrounds such as sky, trees, building undoubtedly increase the computing cost of policing algorithm, meanwhile, non-road surface region is often accompanied by leaf rolling
Dynamic, light variation etc., these disturbing factors also influence the accuracy of monitoring.Therefore, it is necessary to video pictures high speed will be extracted
Pretreatment of the highway pavement region as video monitoring filters out the background unrelated with road surface, reduces the redundant data in image, carries
High arithmetic speed avoids the interference that image information is brought to later image processing in extraneous areas, improves accuracy of detection.
Highway is defined road surface region and is divided track with the white lane line of clearly constant width, then public at a high speed
Road road surface extracted region can all be attributed to the lane line in positioning road image, first extract the lane line on road surface, then utilize vehicle
Road image segmentation is road surface region and non-road surface region by diatom.
Based on above-mentioned theory, Wang et al. in 2004《Image and Vision computing》On deliver
《Image and Vision computing》Assuming that the road in monitoring scene is parallel, it is fitted with B-snake curves
Lane line, with respect to Splines Interpolation Curve, this method causes the curve of generation to approach as possible rather than by interpolation point, the song fitted
Line is flexibly smooth but needs multiple Hough transformation, is unfavorable for control algolithm calculation amount；Jung et al. in 2005《Image
and Vision Computing》On deliver《Lane Following and Lane Departure Using a
Linear-parabolic Model》It is fitted using subregional lane line, road area nearby is fitted using linear model
Lane line, the road area of distant place is fitted lane line using parabola model, but far and near region is to divide in advance in the algorithm
Alright, adaptivity is poor, and use scope is limited；Lipski et al. in 2008《In Proceedings of IEEE
Southwest Symposium on Image Analysis and Interpretation》It delivers《AFast and
Robust Approach to Lane Marking Detection and Lane Tracking》Calculate the office of pavement image
Portion's histogram extracts the characteristic informations such as color and the road direction of pavement image, though this method is by the shape change impacts of road
It is small, but the change of the illumination condition on road surface, shade covering and the decline of lane line clarity etc. can all influence to detect in monitoring environment
As a result；Lee et al. in 2009《In Proceedings of 4th International Conference on
Computer Sciences and Convergence Information Technology》On deliver《Effective
lane detection and tracking method using statistical modeling of color and
lane edge-orientation》Road surface lane line pixel is obtained using track line color and marginal information and calculates edge letter
The histogram of breath and HSV space colouring information, classifies to pixel each in image with bayesian criterion, extracts lane line picture
Element, and Hough transformation is fitted；Kong et al. in 2010《IEEE Trans on Image Processing》It delivers
《General Road Detection from A Signal Image》The office of pixel is calculated using Gabor filtering
Portion's texture obtains marginal information, carries out Hough transformation, to find and position the road surface ROI region in image, though this method is made an uproar
The influence of sound is small but computation complexity is excessively high.
The less model parameter of method demand solution based on model in the above method, calculation amount is small, to noise have compared with
Strong robustness, but selection and method for solving of the correctness in road surface region dependent on model are extracted, based on lane line feature
Method extraction road surface region correctness and the selection of road surface characteristic have direct relation, and highway has depositing for bend
Influence left and right driveway partition accuracy under the scene directly is being fitted within into driveway line on entire picture.
Invention content
The technical problems to be solved by the invention are to provide the express highway pavement detection method based on image procossing, pass through
Using subregion Line segment detection, since bending degree is smooth at expressway bend, the lane line or road edge of small fragment are near
Straightway is similar to, therefore subregion Line segment detection is applicable in except the pavement detection of highway straight way scene, is also added to bend
The adaptability of scene.
To achieve the above object, the present invention provides following technical solution：
A kind of express highway pavement detection method based on image procossing, includes the following steps：
An at least width video frame images for continuous acquisition video file, and obtained according to an at least width video frame images
Target video frame image, wherein, the pixel value of any one pixel in the target video frame image is described at least one
The average value of the grey scale pixel value of corresponding pixel points in width video frame images；
Edge detection is carried out to the target video frame image, obtains the edge image for including road edge pixel point；
The edge image is scanned, obtains road area, and the road area laterally divide and obtains subregion,
And each sub-regions are detected using probability Hough transformation, obtain road edge line segment；
All edge line segments in the subregion on the road area top solve vanishing point and according to each height
The straight line of maximum slope and the straight line of slope minimum determine the centre of each non-bottom end subregion with the presence or absence of intersection point in region
Control point and the boundary point of lowermost end subregion；
According to the intermediate control point, the boundary point, the vanishing point, left and right track edge line is drawn, and according to described
Vehicle determines road surface region to edge line.
Optionally, before described the step of carrying out image filtering processing to the target video frame image, the method
It further includes：
Image filtering processing is carried out to the target video frame image；
It is described that edge detection is carried out to the target video frame image, obtain the edge image for including road edge pixel point
The step of, including：
Edge detection is carried out to the target video frame image after filtering process, obtains the edge for including road edge pixel point
Image.
Optionally, it is described that edge detection is carried out to the target video frame image, it obtains comprising road edge pixel point
The step of edge image, including：
Edge detection is carried out to the target video frame image using Canny edge detection operators, is obtained comprising road roadside
The edge image of edge pixel.
Optionally, the scanning edge image obtains road area, and lateral division is carried out to the road area
The step of obtaining subregion, and each sub-regions are detected using probability Hough transformation, obtaining road edge line segment, packet
It includes：
The edge image is from top to bottom progressively scanned, until there is white pixel point in target line, by the target line
It is confirmed as road area in following region；
The road area is carried out laterally to be divided into third molecular domains, and using probability Hough transformation to each height
Region is detected, and obtains road edge line segment.
Optionally, it is described that each sub-regions are detected using probability Hough transformation, obtain road edge line segment
Step, including：
S1：For each sub-regions, marginal point rendering parameter space curve is randomly selected；
S2：Judge that marginal point concentrates whether the curve quantity for meeting at target point reaches predetermined threshold value, returns to S1, such as if not
Fruit is to perform S3；
S3：By distance between marginal point be less than pre-determined distance point be linked to be line segment, and calculate the length of line segment, and line segment
Length be more than length threshold when be determined as edge line segment.
Optionally, the method further includes：
Judge whether the edge line segment is only present in a sub-regions, if so, deleting the edge line segment.
Optionally, all edge line segments in the subregion according to the road area top solve the formula of vanishing point
It is embodied as：
Vanishing point in the region on top in detection image, to all straight line l in apex zonei, i=1 ... k, it calculates
Straight line liWith vanishing point vpDistance d, solved using random sample consensus algorithm iteration, obtain the vanishing point v of imagep, formula is such as
Under：
Wherein, liFor i-th edge line segment in the subregion on top, in total with k edge line segment, d liPlace
The distance of straight line and vanishing point.
Optionally, described whether there is according to the straight line of maximum slope in each sub-regions and the straight line of slope minimum is handed over
Point determine each non-bottom end subregion intermediate control point and lowermost end subregion boundary point the step of, including：
Determine the side of the first straight line and slope minimum where the edge line segment of maximum slope in the subregion on the top
Second straight line where edge line segment, third straight line and slope in middle sub-field where the edge line segment of maximum slope are minimum
The 5th straight line and slope in the 4th straight line, bottom subregion where edge line segment where the edge line segment of maximum slope is minimum
Edge line segment where the 6th straight line；
When it is not two that first straight line and third straight line, second straight line are with intersection point existing for the 4th straight line, by described the
The intersection point of the coboundary of one straight line, the second straight line and the middle sub-field is controlled as the intermediate of left and right track edge line
Point, if first straight line and the 5th straight line, second straight line with intersection point existing for the 6th straight line are not two, by first straight line,
Boundary point of the intersection point of second branch line and bottom subregion as left and right track edge line, otherwise, by the 5th straight line, the 6th straight line
With the boundary point of the intersection point of bottom subregion as left and right track edge line；
In first straight line and third straight line, second straight line and when intersection point is two existing for the 4th straight line, then by two friendships
Intermediate control point of the point as left and right track edge line, and in the third straight line and the 5th straight line, the 4th straight line
With the 6th straight line there are during two intersection points, using two nodes as the boundary point of left and right track edge line, otherwise, by institute
Third straight line and the 4th straight line are stated with the intersection point of the image boundary of the bottom subregion as left and right track edge line
Boundary point.
Optionally, it is described according to the intermediate control point, the boundary point, the vanishing point, left and right track edge line is drawn,
And the step of road surface region is determined according to the vehicle to edge line, including：
According to the intermediate control point, the boundary point, the vanishing point, left and right vehicle is drawn using cubic spline functions
Road edge line, and road surface region is determined according to the vehicle to edge line.
Optionally, the method further includes：
According to the road surface region, watershed point is found out, and subregion clicks through the watershed using watershed algorithm
Row probability Hough transformation completes the detection of middle lane cut-off rule using cubic spline interpolation.
Express highway pavement detection method provided by the invention based on image procossing, has the advantages that：
1st, by the express highway pavement detection method provided by the invention based on image procossing, to road static Background
As carrying out differential filtering, using the intensity profile feature of lane line horizontal pixel in pavement image, it can effectively weaken mileage chart
The luminance information of non-rice habitats area pixel point as in, but can well in very prominent image lane line edge pixel；
2nd, it by the express highway pavement detection method provided by the invention based on image procossing, is examined with subregion line segment
It surveys, since bending degree is smooth at expressway bend, the lane line or road edge of small fragment are similar to straightway, therefore subregion
Domain Line segment detection is applicable in except the pavement detection of highway straight way scene, also adds the adaptability to bend scene；
3rd, it is consistent using random sample by the express highway pavement detection method provided by the invention based on image procossing
Property algorithm extraction pavement image vanishing point, the method for avoiding the detection vanishing point of mean value method in straight way scene do not adapt to bend field
The problem of scape, improves robustness of the vanishing point estimation to scene.
Description of the drawings
Fig. 1 is the schematic diagram of the express highway pavement detection method based on image procossing of the embodiment of the present invention；
Fig. 2 is the corresponding gray-scale map of target video frame image of the embodiment of the present invention；
Fig. 3 is the gray-scale map after Fig. 2 of the embodiment of the present invention is filtered；
Fig. 4 is the detection result figure of the edge pixel of the embodiment of the present invention；
Fig. 5 is the flow chart of the probability Hough transformation detection road edge circuit of the embodiment of the present invention；
Fig. 6 is the schematic diagram for the line segment that Fig. 5 is detected；
Fig. 7 is the line segment schematic diagram after bicycle diatom is rejected；
Fig. 8 is the vanishing point position view detected in the embodiment of the present invention；
Fig. 9 is the distribution schematic diagram of vanishing point, intermediate control point and boundary point；
Figure 10 is the left and right track edge line schematic diagram drawn；
The road surface area schematic that Figure 11 is made of Figure 10；
Figure 12 is the left and right driveway partition line obtained using watershed algorithm；
Figure 13 is the middle lane cut-off rule that cubic spline interpolation is used on the basis of Figure 12；
Figure 14 a be three lane lines proposed based on the embodiment of the present invention,
White area in Figure 14 b is the road surface region based on left and right track line drawing；
White area in Figure 14 c is the region where the left side road surface based on three track line drawings；
White area in Figure 14 d is the region where the right road surface based on three vehicle track line drawings.
Specific embodiment
Purpose, technical scheme and advantage to make invention are of greater clarity, below by attached drawing and embodiment, to this
Inventive technique scheme is further elaborated.However, it should be understood that specific embodiment described herein is only used to solve
Technical solution of the present invention is released, is not intended to restrict the invention the range of technical solution.
To solve prior art problem, the present invention provides embodiments as shown in Figure 1.
A kind of express highway pavement detection method based on image procossing of the present invention：Lane line edge pixel is first passed through to carry
It takes, after background image extraction is carried out to the highway image of reading matter then progress differential filtering extracts road edge, Ran Houtong
It crosses and after being divided to image-region each sub-regions is carried out with probability Hough (Hough) detection, then pass through vanishing point and estimate to carry out
Then left and right edges line drawing carries out left and right driveway partition so as to carry out road surface extracted region.Wherein, RANSAC Random
The abbreviation of Sample Consensus, it is the sample data set for including abnormal data according to one group, calculates the mathematics of data
Model parameter obtains the algorithm of effective sample data.RANSAC algorithms are frequently used in computer vision.For example, in stereopsis
The calculating of the match point and fundamental matrix of a pair of of camera is solved the problems, such as in feel field simultaneously.
A kind of expressway surface based on lane line provided in an embodiment of the present invention detects detailed process and is：
1st, lane line edge pixel extracts：
(11) the preceding N frames image I of continuous acquisition video sequence1I2…Ii…IN, illustratively, setting N=20 pictures correspond to picture
Grey scale pixel value B (x, y) of the element value average value as highway static background image, calculation formula are as follows：
Wherein, Ii(x, y) represents the i-th frame image corresponding grey scale pixel value at pixel (x, y), and B (x, y) is background
Corresponding grey scale pixel value at pixel (x, y) in image, i.e., target video frame image is in the corresponding grey scale pixel value in the shop
Average value, obtained image are as shown in Figure 2.
(12) the grey scale change information for the lane line horizontal pixel in effectively prominent image, is presented white with reference to lane line
Color and there is certain width, and the characteristics of the road surface brightness value on side is low, the gray value B (x, y) of each pixel to background image
With the gray value of the neighborhood territory pixel of its left and right distance τ difference is carried out, with their difference and filtered image is represented, such as formula
(2):
Bf(x, y)=2B (x, y)-B (x- τ, y)-B (x+ τ, y)-| B (x- τ, y)-B (x+ τ, y) | (2)
In formula, the step-length of τ wave filters, according to the width in highway lane line region, τ=10 here.By formula (2)
Filtered image is as shown in Figure 3.
(13) edge detection is carried out to the target video frame image after filtered
Edge detection, corresponding side are carried out to target video frame shown in Fig. 3 using Canny operators in the embodiment of the present invention
Edge pixel map is as shown in Figure 4.Canny operators ask marginal point specific algorithm step as follows：With Gaussian filter smoothed image；With one
Rank local derviation finite difference formulations gradient magnitude and direction；Non-maxima suppression is carried out to gradient magnitude；It is detected with dual threashold value-based algorithm
With connection edge.Concrete implementation process is the prior art, and the embodiment of the present invention does not repeat it herein.
(2) subregion detection lane line, from top to bottom progressively scanned filtered target video frame image first, until
There is white pixel point in the row, and the image-region more than row is divided into sky areas, is not detected, the region below the row
It is determined as road area, lateral trisection region division is carried out to the corresponding image of the road area, then utilizes probability Hough
Change brings subregion detection road edge line segment, as shown in Figure 5 for the flow chart of the probability Hough transformation of each sub-regions：
(21) for each sub-regions, judge group of edge points in subregion into edge point set whether be empty, if not
A marginal point is then randomly selected in subregion edge image for sky, judges the point whether in upper one cycle probability Hough transformation
In be marked as point in target line, if it is, continue to randomly select a marginal point in remaining marginal point, until
All marginal points have all been extracted；
(22) these are not located to target line point and is transformed into the parameter space defined by parameter ρ and θ, wherein, ρ represents figure
To the vertical range of straight line, θ represents the vertical line and the angle of X-axis, calculates the marginal point on all directions θ origin in image space
ρ, and corresponding A (ρ, θ) is cumulative and calculates, A (ρ, θ) represent rectilinear point in this direction number；
(23) it is concentrated from marginal point and deletes taken point, illustratively, if cumulative and A (ρ, θ) is then walked more than 80
It is rapid 4) otherwise to return to step (22)；
(24) according to the corresponding parameters of A (ρ, θ) in (23), such as distance between points less than 20 if be linked to be line segment, then
These points are deleted, and record the starting point and ending point of the line segment, otherwise return to (21)；
(25) length of line segment is calculated, if greater than length threshold 100, then the line segment output for being considered as, and delete
Corresponding marginal point is concentrated at edge, otherwise returns to (21)；
(26) (21) to (25) are repeated, the straightway that the probability transformation that the closed grey line segment in Fig. 6 is Fig. 4 extracts.
(3) road surface extracted region process：
(31) the lane line line segment of grey in Fig. 6 is judged, if the line segment of detection is only present in certain sub-regions,
Its straight line in detection is concentrated and deleted by then its short lane line or flase drop for road surface region division track, as shown in fig. 7, through
Step judgement is crossed, will respectively detect that a short lane line is rejected in bottom zone in Fig. 6 and apex zone, detailed process
It is as follows：
(32) to all line segment l in apex zonei, i=1 ... k, calculating straight line liWith vanishing point vpDistance d, using with
Press proof this consistency algorithm iterative solution formula (3), obtains final vanishing point vp, such as the white dot in Fig. 8；
(33) it sorts in three sub-regions to the straight line in (31) into line slope, determines slope in the subregion on the top
First straight line where maximum edge line segment and the second straight line where the edge line segment of slope minimum, it is oblique in middle sub-field
Third straight line where the edge line segment of rate maximum and the 4th straight line where the edge line segment of slope minimum, in the subregion of bottom
The 5th straight line where the edge line segment of maximum slope and the 6th straight line where the edge line segment of slope minimum；
When it is not two that first straight line and third straight line, second straight line are with intersection point existing for the 4th straight line, by described the
The intersection point of the coboundary of one straight line, the second straight line and the middle sub-field is controlled as the intermediate of left and right track edge line
Point, if first straight line and the 5th straight line, second straight line with intersection point existing for the 6th straight line are not two, by first straight line,
Boundary point of the intersection point of second branch line and bottom subregion as left and right track edge line, otherwise, by the 5th straight line, the 6th straight line
With the boundary point of the intersection point of bottom subregion as left and right track edge line；
In first straight line and third straight line, second straight line and when intersection point is two existing for the 4th straight line, then by two friendships
Intermediate control point of the point as left and right track edge line, and in the third straight line and the 5th straight line, the 4th straight line
With the 6th straight line there are during two intersection points, using two nodes as the boundary point of left and right track edge line, otherwise, by institute
State third straight line and the 4th straight line and boundary of the intersection point of the image boundary of bottom subregion as left and right track edge line
Point.Such as Fig. 9, finally the white dot of borderline two of top area in determining intermediate control point such as Fig. 9, final to determine
The white dot of image right margin and borderline two of bottom end in boundary point such as Fig. 9；
(34) based on the vanishing point in (32), intermediate control point and boundary point in (33), using cubic spline functions
Left and right track edge line is drawn, such as the left and right track edge line of two black in Figure 10, the area that two track edge lines define
Domain is road surface region, as shown in figure 11.
(4) left and right driveway partition
The process of left and right driveway partition：Mask label figure, the marked region of left lane are obtained in road area image
Value for 1, the value of the marked region in the right track is 2, and the value in unmarked region is 0, and watershed point is found with watershed algorithm,
Such as the black curve in Figure 12, and subregion carries out probability Hough transformation to these points, based in method extraction in step (33)
Between point and marginal point, using cubic spline interpolation complete middle lane cut-off rule detection, such as the centre of black in Figure 13
Track cut-off rule based on left and right track edge line and middle lane cut-off rule, marks off left and right track region.
The detection scene of this example is the highway that there is bend in distal end, and implementation result is as shown in figure 14, in Figure 14 (a)
Three lane lines of black are the lane line of the method for the present invention extraction, and the white area in Figure 14 (b) is based on left and right lane line
The road surface region of extraction, the white area in Figure 14 (c) are the left side road surface region based on three track line drawings,
White area in Figure 14 (d) is the right road surface region based on three track line drawings, it can be seen that calculation of the invention
Method can be extracted to the road surface region in bend section and accurately be partitioned into left and right track well.
In addition, it should be understood that although this specification is described in terms of embodiments, but not each embodiment is only wrapped
Containing an independent technical solution, this description of the specification is merely for the sake of clarity, and those skilled in the art should
It considers the specification as a whole, the technical solutions in each embodiment can also be properly combined, forms those skilled in the art
The other embodiment being appreciated that.
Claims (8)
1. a kind of express highway pavement detection method based on image procossing, which is characterized in that including：
An at least width video frame images for continuous acquisition video file, and obtain target according to an at least width video frame images
Video frame images, wherein, the pixel value of any one pixel in the target video frame image is that an at least width regards
The average value of the grey scale pixel value of corresponding pixel points in frequency frame image；
Edge detection is carried out to the target video frame image, obtains the edge image for including road edge pixel point；
The edge image is scanned, obtains road area, and the road area laterally divide and obtains subregion, and profit
Each sub-regions are detected with probability Hough transformation, obtain road edge line segment；
All edge line segments in the subregion on the road area top solve vanishing point and according to each sub-regions
The straight line of straight line and the slope minimum of middle maximum slope determines that the intermediate of each non-bottom end subregion controls with the presence or absence of intersection point
The boundary point of point and lowermost end subregion；
According to the intermediate control point, the boundary point, the vanishing point, left and right track edge line is drawn, and according to the track
Edge line determines road surface region；
The scanning edge image obtains road area, and the road area laterally divide and obtains subregion,
And the step of each sub-regions are detected using probability Hough transformation, obtain road edge line segment, including：
The edge image is from top to bottom progressively scanned, until there is white pixel point in target line, below the target line
Region be confirmed as road area；
The road area is carried out laterally to be divided into third molecular domains, and using probability Hough transformation to each sub-regions
It is detected, obtains road edge line segment；
The step of utilization probability Hough transformation is detected each sub-regions, obtains road edge line segment, including：
S1：For each sub-regions, marginal point rendering parameter space curve is randomly selected；
S2：Judge that marginal point concentrates whether the curve quantity for meeting at target point reaches predetermined threshold value, returns to S1 if not, if
Perform S3；
S3：By distance between marginal point be less than pre-determined distance point be linked to be line segment, and calculate the length of line segment, and line segment length
Degree is determined as edge line segment when being more than length threshold.
2. a kind of express highway pavement detection method based on image procossing according to claim 1, it is characterised in that：Institute
Before stating the step of carrying out edge detection to the target video frame image, the method further includes：
Image filtering processing is carried out to the target video frame image；
It is described that edge detection is carried out to the target video frame image, obtain the step of the edge image comprising road edge pixel point
Suddenly, including：
Edge detection is carried out to the target video frame image after filtering process, obtains the edge graph for including road edge pixel point
Picture.
A kind of 3. express highway pavement detection method based on image procossing according to claim 2, which is characterized in that institute
The step of stating and edge detection is carried out to the target video frame image, obtaining the edge image comprising road edge pixel point, packet
It includes：
Edge detection is carried out to the target video frame image using Canny edge detection operators, is obtained comprising road edge picture
The edge image of vegetarian refreshments.
A kind of 4. express highway pavement detection method based on image procossing according to claim 5, which is characterized in that institute
The method of stating further includes：
Judge whether the edge line segment is only present in a sub-regions, if so, deleting the edge line segment.
5. a kind of express highway pavement detection method based on image procossing according to claim 1 or 4, feature exist
In the formula that all edge line segments in the subregion according to the road area top solve vanishing point is embodied as：
Vanishing point in the region on top in detection image, to all straight line l in apex zonei, i=1 ... k, calculate straight line
liWith vanishing point vpDistance d, solved using random sample consensus algorithm iteration, obtain the vanishing point v of imagep, formula is as follows：
Wherein, liFor i-th edge line segment in the subregion on top, in total with k edge line segment, d liThe straight line at place
With the distance of vanishing point.
A kind of 6. express highway pavement detection method based on image procossing according to claim 4, which is characterized in that institute
It states and each non-bottom is determined with the presence or absence of intersection point according to the straight line of maximum slope in each sub-regions and the straight line of slope minimum
The step of boundary point of the intermediate control point of terminal area and lowermost end subregion, including：
Determine the edge line of the first straight line and slope minimum where the edge line segment of maximum slope in the subregion on the top
Second straight line where section, the edge of third straight line and slope minimum in middle sub-field where the edge line segment of maximum slope
The side of the 5th straight line and slope minimum in the 4th straight line, bottom subregion where line segment where the edge line segment of maximum slope
The 6th straight line where edge line segment；
It is straight by described first when it is not two that first straight line and third straight line, second straight line are with intersection point existing for the 4th straight line
Intermediate control point of the intersection point of line, the second straight line and the coboundary of the middle sub-field as left and right track edge line,
If first straight line and the 5th straight line, second straight line and intersection point existing for the 6th straight line are not two, by first straight line, second
Boundary point of the intersection point of branch line and bottom subregion as left and right track edge line, otherwise, by the 5th straight line, the 6th straight line and bottom
Boundary point of the intersection point of portion's subregion as left and right track edge line；
In first straight line and third straight line, second straight line and when intersection point is two existing for the 4th straight line, then two intersection points are made
For the intermediate control point of left and right track edge line, and in the third straight line and the 5th straight line, the 4th straight line and institute
The 6th straight line is stated there are during two intersection points, using two nodes as the boundary point of left and right track edge line, otherwise, by described
Boundary of the intersection point of the image boundary of three straight lines and the 4th straight line and the bottom subregion as left and right track edge line
Point.
A kind of 7. express highway pavement detection method based on image procossing according to claim 1, which is characterized in that institute
It states according to the intermediate control point, the boundary point, the vanishing point, draws left and right track edge line, and according to the vehicle to side
Edge line determines the step of road surface region, including：
According to the intermediate control point, the boundary point, the vanishing point, left and right lane side is drawn using cubic spline functions
Edge line, and road surface region is determined according to the vehicle to edge line.
A kind of 8. express highway pavement detection method based on image procossing according to claim 2, which is characterized in that institute
The method of stating further includes：According to the road surface region, watershed point is found out, and subregion is to the watershed using watershed algorithm
Point carries out probability Hough transformation, and the detection of middle lane cut-off rule is completed using cubic spline interpolation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711480274.5A CN108171695A (en) | 2017-12-29 | 2017-12-29 | A kind of express highway pavement detection method based on image procossing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711480274.5A CN108171695A (en) | 2017-12-29 | 2017-12-29 | A kind of express highway pavement detection method based on image procossing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108171695A true CN108171695A (en) | 2018-06-15 |
Family
ID=62516351
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711480274.5A Pending CN108171695A (en) | 2017-12-29 | 2017-12-29 | A kind of express highway pavement detection method based on image procossing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108171695A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544577A (en) * | 2018-11-27 | 2019-03-29 | 辽宁工程技术大学 | A kind of improvement lines detection method based on marginal point marshalling |
| CN109785463A (en) * | 2019-01-25 | 2019-05-21 | 芜湖智久机器人有限公司 | A kind of carriage fault detection system and method based on speed |
| CN110021167A (en) * | 2019-04-09 | 2019-07-16 | 杭州路灵信息技术有限公司 | A kind of highway cloud inspection system and data processing method |
| CN110827393A (en) * | 2019-07-17 | 2020-02-21 | 当家移动绿色互联网技术集团有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN110969576A (en) * | 2019-11-13 | 2020-04-07 | 同济大学 | Highway pavement image splicing method based on roadside PTZ camera |
| CN111723734A (en) * | 2020-06-19 | 2020-09-29 | 哈尔滨理工大学 | Hole site identification method based on machine vision |
| CN113033395A (en) * | 2021-03-25 | 2021-06-25 | 太原科技大学 | Drivable region segmentation method based on DeFCN and vanishing point edge detection |
| CN113643272A (en) * | 2021-08-24 | 2021-11-12 | 凌云光技术股份有限公司 | Target positioning modeling method |
| CN113688812A (en) * | 2021-10-27 | 2021-11-23 | 之江实验室 | Expressway projectile detection method based on background modeling and classification network |
| CN114862753A (en) * | 2022-03-17 | 2022-08-05 | 北京梦诚科技有限公司 | Automatic high-precision table correction method and system |
| CN116580032A (en) * | 2023-07-14 | 2023-08-11 | 青岛西海岸城市建设集团有限公司 | Quality monitoring method for road construction |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5710554A (en) * | 1996-02-01 | 1998-01-20 | Pettler; Peter R. | Pavement ice detector |
| WO2009042933A1 (en) * | 2007-09-28 | 2009-04-02 | The Mainz Group, Llc | Photogrammetric networks for positional accuracy and ray mapping |
| CN105760812A (en) * | 2016-01-15 | 2016-07-13 | 北京工业大学 | Hough transform-based lane line detection method |
| CN106485663A (en) * | 2015-08-26 | 2017-03-08 | 腾讯科技（深圳）有限公司 | A kind of lane line image enchancing method and system |
| CN106682563A (en) * | 2015-11-05 | 2017-05-17 | 腾讯科技（深圳）有限公司 | Lane line detection self-adaptive adjusting method and device |
-
2017
- 2017-12-29 CN CN201711480274.5A patent/CN108171695A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5710554A (en) * | 1996-02-01 | 1998-01-20 | Pettler; Peter R. | Pavement ice detector |
| WO2009042933A1 (en) * | 2007-09-28 | 2009-04-02 | The Mainz Group, Llc | Photogrammetric networks for positional accuracy and ray mapping |
| CN106485663A (en) * | 2015-08-26 | 2017-03-08 | 腾讯科技（深圳）有限公司 | A kind of lane line image enchancing method and system |
| CN106682563A (en) * | 2015-11-05 | 2017-05-17 | 腾讯科技（深圳）有限公司 | Lane line detection self-adaptive adjusting method and device |
| CN105760812A (en) * | 2016-01-15 | 2016-07-13 | 北京工业大学 | Hough transform-based lane line detection method |
Non-Patent Citations (2)
| Title |
|---|
| 李畅: "《由近景影像反演建筑立面三维模型研究》", 31 July 2016, 武汉：华中师范大学出版社 * |
| 隆迪: "高速公路弯道场景的路面区域提取", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109544577A (en) * | 2018-11-27 | 2019-03-29 | 辽宁工程技术大学 | A kind of improvement lines detection method based on marginal point marshalling |
| CN109544577B (en) * | 2018-11-27 | 2022-10-14 | 辽宁工程技术大学 | Improved straight line extraction method based on edge point grouping |
| CN109785463A (en) * | 2019-01-25 | 2019-05-21 | 芜湖智久机器人有限公司 | A kind of carriage fault detection system and method based on speed |
| CN110021167A (en) * | 2019-04-09 | 2019-07-16 | 杭州路灵信息技术有限公司 | A kind of highway cloud inspection system and data processing method |
| CN110827393A (en) * | 2019-07-17 | 2020-02-21 | 当家移动绿色互联网技术集团有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN110827393B (en) * | 2019-07-17 | 2020-07-31 | 当家移动绿色互联网技术集团有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN110969576B (en) * | 2019-11-13 | 2021-09-03 | 同济大学 | Highway pavement image splicing method based on roadside PTZ camera |
| CN110969576A (en) * | 2019-11-13 | 2020-04-07 | 同济大学 | Highway pavement image splicing method based on roadside PTZ camera |
| CN111723734A (en) * | 2020-06-19 | 2020-09-29 | 哈尔滨理工大学 | Hole site identification method based on machine vision |
| CN111723734B (en) * | 2020-06-19 | 2022-02-18 | 哈尔滨理工大学 | Hole site identification method based on machine vision |
| CN113033395A (en) * | 2021-03-25 | 2021-06-25 | 太原科技大学 | Drivable region segmentation method based on DeFCN and vanishing point edge detection |
| CN113643272A (en) * | 2021-08-24 | 2021-11-12 | 凌云光技术股份有限公司 | Target positioning modeling method |
| CN113688812A (en) * | 2021-10-27 | 2021-11-23 | 之江实验室 | Expressway projectile detection method based on background modeling and classification network |
| CN114862753A (en) * | 2022-03-17 | 2022-08-05 | 北京梦诚科技有限公司 | Automatic high-precision table correction method and system |
| CN116580032A (en) * | 2023-07-14 | 2023-08-11 | 青岛西海岸城市建设集团有限公司 | Quality monitoring method for road construction |
| CN116580032B (en) * | 2023-07-14 | 2023-09-26 | 青岛西海岸城市建设集团有限公司 | Quality monitoring method for road construction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108280450A (en) | A kind of express highway pavement detection method based on lane line | |
| CN108171695A (en) | A kind of express highway pavement detection method based on image procossing | |
| Cheng et al. | Extraction and classification of road markings using mobile laser scanning point clouds | |
| CN110569704B (en) | Multi-strategy self-adaptive lane line detection method based on stereoscopic vision | |
| CN107045629B (en) | Multi-lane line detection method | |
| Son et al. | Real-time illumination invariant lane detection for lane departure warning system | |
| CN105260699B (en) | A kind of processing method and processing device of lane line data | |
| CN105678285B (en) | A kind of adaptive road birds-eye view transform method and road track detection method | |
| WO2015010451A1 (en) | Method for road detection from one image | |
| CN109784344A (en) | A kind of non-targeted filtering method of image for ground level mark identification | |
| CN107862290A (en) | Method for detecting lane lines and system | |
| Zhang et al. | Semi-automatic road tracking by template matching and distance transformation in urban areas | |
| Li et al. | Road detection algorithm for autonomous navigation systems based on dark channel prior and vanishing point in complex road scenes | |
| CN104700071B (en) | A kind of extracting method of panorama sketch road profile | |
| Yaghoobi Ershadi et al. | Robust vehicle detection in different weather conditions: Using MIPM | |
| CN103617412A (en) | Real-time lane line detection method | |
| CN110197173B (en) | Road edge detection method based on binocular vision | |
| CN113239733B (en) | Multi-lane line detection method | |
| CN107705301A (en) | A kind of highway graticule damage testing method based on unmanned plane highway map picture | |
| CN105809149A (en) | Lane line detection method based on straight lines with maximum length | |
| Liu et al. | Development of a vision-based driver assistance system with lane departure warning and forward collision warning functions | |
| CN106558051A (en) | A kind of improved method for detecting road from single image | |
| CN110414385A (en) | A kind of method for detecting lane lines and system based on homography conversion and characteristic window | |
| CN105809673A (en) | SURF (Speeded-Up Robust Features) algorithm and maximal similarity region merging based video foreground segmentation method | |
| CN117094914A (en) | Smart city road monitoring system based on computer vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180615 |