CN105137412A - Accurate fitting method of line segment features in 2D laser radar distance image - Google Patents

Accurate fitting method of line segment features in 2D laser radar distance image Download PDF

Info

Publication number
CN105137412A
CN105137412A CN201510511455.4A CN201510511455A CN105137412A CN 105137412 A CN105137412 A CN 105137412A CN 201510511455 A CN201510511455 A CN 201510511455A CN 105137412 A CN105137412 A CN 105137412A
Authority
CN
China
Prior art keywords
data
line
line segment
laser radar
segment
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.)
Granted
Application number
CN201510511455.4A
Other languages
Chinese (zh)
Other versions
CN105137412B (en
Inventor
赵敏
孙棣华
熊星
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing Kezhiyuan Technology Co ltd
Original Assignee
Chongqing University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chongqing University filed Critical Chongqing University
Priority to CN201510511455.4A priority Critical patent/CN105137412B/en
Publication of CN105137412A publication Critical patent/CN105137412A/en
Application granted granted Critical
Publication of CN105137412B publication Critical patent/CN105137412B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses an accurate fitting method of line segment features in a 2D laser radar distance image. The method comprises acquiring the laser radar distance image data to perform data pre processing; converting the laser radar distance image data into the polar coordinate model data; performing adaptive threshold area segmentation of the polar coordinate model data to form the area data; performing line segmentation; and at last merging the line segment data and outputting the merged line segment data, and taking the output merged line segment data as the extracted line segment data of the laser radar distance image. Based on the traditional Split-and-Merge algorithm, the accurate fitting method of line segment features in a 2D laser radar distance image utilizes two-stage segmentation at first, and then can obtain the line segment information in the laser radar distance image through merging and fitting, and utilizes the adaptive dynamic method to select segmented threshold values and optimizes the threshold values on the merging criterion so that the defect that a traditional method needs to adjust the parameters under different scenes is overcome; the line segment features in the laser radar data image can be extracted quickly and accurately; and the accurate fitting method of line segment features is suitable for different indoor scenes and can guarantee the accuracy of fitting.

Description

A kind of 2D laser radar range image middle conductor feature Accurate Curve-fitting method
Technical field
The present invention relates to sensing detection field, particularly a kind of 2D laser radar range image middle conductor feature Accurate Curve-fitting method, be applicable to the Accurate Curve-fitting to 2D laser radar range image middle conductor feature under doors structure scene.
Background technology
The primary selection that 2D laser radar is high with its precision, investigative range is wide, the anti-interference feature having cost performance such as strong, moderate becomes mobile robot's independent navigation under circumstances not known.Laser radar is using laser as signal source, run into testee back reflection by laser instrument with the pulse laser that certain angular emission goes out and return receiver, thus the distance of testee is measured according to the time interval (TOF) of transmitting-receiving, wherein 2D laser radar is by certain limit detection range information a plane.
Under the scene of doors structure, the range image of laser radar sensing external environment often presents the feature of line segment.Such as under corridor environment, image presents two parallel segments, and under the environment of corner, image presents two orthogonal line segments.Only have the range information extracted quickly and accurately in image, modeling could be carried out to environment.
The range information how obtained according to laser radar simulates line segment information accurately, that mobile robot is to the key detected under circumstances not known, at present, existing fit line phase method often adopts the line segments extraction method based on 2D laser radar, mainly contains: Split-and-Merge algorithm, Incremental algorithm, Hough transformation algorithm etc.; Wherein, Split-and-Merge algorithm speed is fast, but the effect extracted is chosen threshold value and to split the dependence of merging criterion comparatively large, needs to select suitable parameter under different scenes.Incremental algorithm speed is fast, and complexity is low, but this algorithm is mainly applicable to the simple scene of structure, and the fitting effect for complex scene downcrossings straight line is bad.Hough transformation has good noise immunity.But calculated amount is very large, should not adopt in the independent navigation higher to requirement of real-time, and threshold value choose also more difficult.
Because above-mentioned algorithm exists the problem that adaptability is bad, computation complexity is higher, precision is not high, be difficult to adapt to complex scene.Therefore, be badly in need of a kind of both having there is good adaptability, the laser radar line segments extraction method of rapidity, accuracy can be ensured again.
Summary of the invention
In view of this, technical matters to be solved by this invention is to provide a kind of 2D laser radar range image middle conductor feature Accurate Curve-fitting method, is applicable to the Accurate Curve-fitting to 2D laser radar range image middle conductor feature under doors structure scene.
The object of the present invention is achieved like this:
A kind of 2D laser radar range image middle conductor feature Accurate Curve-fitting method provided by the invention, comprises the following steps:
S1: obtain laser radar range image data and carry out data prediction;
S2: be Polar Coordinate Model data by range image data transformations after data prediction;
S3: to Polar Coordinate Model data acquisition adaptive threshold region segmentation forming region data;
S4: line segment segmentation is carried out to each area data and forms cut-off rule segment data;
S5: line segment merging is carried out to the segment data in cut-off rule segment data;
S6: until all cut-off rule segment datas merge complete and export merging segment data, described merging segment data is as the extraction segment data of laser radar range image.
Further, in described S1, data prediction filters random disturbance data in range image data by choosing middle position value filtering method.
Further, in described S3, adaptive threshold region segmentation concrete steps are as follows:
S31: choose current data from Polar Coordinate Model data;
S32: judge whether the range data in current data is greater than zero, if so, then using current data sequence number as number label flag;
S33: if not, then returning step S31, to choose next Polar Coordinate Model data be current data;
S34: the front and back changing value calculating current data and upper data; Calculate the zero data number contained between current data and upper non-vanishing Polar Coordinate Model data;
S35: judge whether front and back changing value is greater than first threshold theta1, or whether zero data number is greater than Second Threshold theta2, if so, then current data is divided into m+1 cut zone data, and returning step S31, to choose next Polar Coordinate Model data be current data;
S36: if not, be then divided into m cut zone data by current data; And returning step S31, to choose next Polar Coordinate Model data be current data;
S37: to the last Polar Coordinate Model data;
S38: the number calculating contained data in each cut zone, is less than being considered as interference region and giving up of predetermined threshold value by number.
Further, the line segment segmentation in described S4 comprises the following steps:
S41: convert the polar data of the area data in each cut zone to rectangular coordinate data;
S42: select two area datas in cut zone and go out straight line according to its rectangular coordinate data fitting;
S43: calculate the point of this area data other data interior to this straight line ultimate range, and calculate ultimate range;
These data if so, are then arranged at the first straight-line data region by S44: judge whether ultimate range is greater than adaptive threshold; If not, then these data are arranged at the second straight-line data region;
Wherein, described adaptive threshold according to ultimate range according to following formulae discovery:
t h e t a = ( x k - x k - 1 ) 2 + ( y k - y k - 1 ) 2 + ( x k + 1 - x k ) 2 + ( y k + 1 - y k ) 2 , Wherein, (x k-1, y k-1) represent kth-1 rectangular coordinate data, (x k, y k) represent a kth rectangular coordinate data, (x k+1, y k+1) represent kth+1 rectangular coordinate data;
S45: be cycled to repeat and the first straight-line data region and the second straight-line data region are split, until all Data Segmentations are complete;
S46: the number calculating contained segmentation straight line in each cut zone, is less than being considered as interference region and giving up of predetermined threshold value by number.
Further, described S5 middle conductor merging specifically comprises the following steps:
S51: take Least Square method to carry out fitting a straight line to the coordinate data in cut-off rule segment data, forms straight-line equation y=k*x+b;
S52: select first coordinate data in line segment coordinate data and last coordinate data, and respectively vertical line is done to straight-line equation y=k*x+b, origin coordinates and the end coordinate of this straight-line equation is obtained by the intersection point calculating described vertical line and described straight-line equation;
S53: convert straight-line equation to polar equation;
S54: the differential seat angle absolute value of the adjacent segments in computed segmentation segment data, and the range difference absolute value of adjacent segments in cut-off rule segment data;
S55: judge whether differential seat angle absolute value is less than angle fixed threshold, and whether range difference absolute value is less than distance fixed threshold; If so, then two adjacent segment line segments are carried out merging and form renewal line segment, and replace the line segment before merging to upgrade line segment; If not, then next step is entered;
S56: return step S51 and be cycled to repeat, until all cut-off rule segment datas all calculate complete.
Beneficial effect of the present invention is: the present invention, on the basis of traditional Split-and-Merge algorithm, first adopts two-stage segmentation, then obtains the line segment information in laser radar range image by merging matching.First order segmentation utilizes polar coordinates middle distance and angle information to extract area information, and second level segmentation is then extract line segment information for area information.And choosing of segmentation threshold adopts the dynamic method of self-adaptation, overcome the shortcoming that classic method needs to adjust parameter under different scenes.The criterion that line segment merges is optimized; The present invention can extract the line segment feature in laser radar data image fast and accurately, and the advantage of the method does not need manually to arrange threshold value, is applicable to different indoor scenes, can ensure the accuracy of matching simultaneously.
Accompanying drawing explanation
In order to make the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail, wherein:
The line-fitting method flow diagram that Fig. 1 provides for the embodiment of the present invention;
The region segmentation process flow diagram that Fig. 2 provides for the embodiment of the present invention;
The line segment segmentation process flow diagram that Fig. 3 provides for the embodiment of the present invention;
The line segment that Fig. 4 provides for the embodiment of the present invention merges process flow diagram.
Embodiment
Hereinafter with reference to accompanying drawing, the preferred embodiments of the present invention are described in detail.Should be appreciated that preferred embodiment only in order to the present invention is described, instead of in order to limit the scope of the invention.
Embodiment 1
As shown in Figure 1,2D laser radar range image middle conductor feature Accurate Curve-fitting method provided by the invention, comprises the following steps:
S1: obtain laser radar range image data and carry out data prediction;
S2: be Polar Coordinate Model data by range image data transformations after data prediction; Described Polar Coordinate Model data are made up of angle value and distance value; Described angle value is θ ithe angle that corresponding intersection point and X-axis are formed; Distance value is p icorresponding initial point is to the distance of straight line;
S3: to Polar Coordinate Model data acquisition adaptive threshold region segmentation forming region data;
In described S3, adaptive threshold region segmentation concrete steps are as follows:
S31: choose current data from Polar Coordinate Model data;
S32: judge whether the range data in current data is greater than zero, if so, then using current data sequence number as number label flag;
The flag of the present embodiment for recording the sequence number of last non-zero, when be recycled to next time non-zero time, the sequence number difference of non-zero and last non-zero next time can be calculated;
S33: if not, then returning step S31, to choose next Polar Coordinate Model data be current data;
S34: the front and back changing value calculating current data and upper data; Calculate the zero data number contained between current data and upper non-vanishing Polar Coordinate Model data;
S35: judge whether front and back changing value is greater than first threshold theta1, and whether zero data number is greater than Second Threshold theta2, if so, then current data is divided into m+1 cut zone data, and returning step S31, to choose next Polar Coordinate Model data be current data;
S36: if not, be then divided into m cut zone data by current data; And returning step S31, to choose next Polar Coordinate Model data be current data;
S37: to the last Polar Coordinate Model data;
S38: the number calculating contained data in each cut zone, is less than being considered as interference region and giving up of predetermined threshold value by number.
S4: line segment segmentation is carried out to each area data and forms cut-off rule segment data;
Line segment segmentation in described S4 comprises the following steps:
S41: convert the polar data of the area data in each cut zone to rectangular coordinate data;
S42: select two area datas in cut zone and go out straight line according to its rectangular coordinate data fitting;
S43: calculate the point of this area data other data interior to this straight line ultimate range, and calculate ultimate range;
These data if so, are then arranged at the first straight-line data region by S44: judge whether ultimate range is greater than adaptive threshold; If not, then these data are arranged at the second straight-line data region;
Wherein, described adaptive threshold according to ultimate range according to following formulae discovery:
t h e t a = ( x k - x k - 1 ) 2 + ( y k - y k - 1 ) 2 + ( x k + 1 - x k ) 2 + ( y k + 1 - y k ) 2 , Wherein, theta represents adaptive threshold; (x k-1, y k-1) represent kth-1 rectangular coordinate data, (x k, y k) represent a kth rectangular coordinate data, (x k+1, y k+1) represent kth+1 rectangular coordinate data;
S45: be cycled to repeat and the first straight-line data region and the second straight-line data region are split, until all Data Segmentations are complete;
S46: the number calculating contained segmentation straight line in each cut zone, is less than being considered as interference region and giving up of predetermined threshold value by number.
S5: line segment merging is carried out to the segment data in cut-off rule segment data;
Described S5 middle conductor merging specifically comprises the following steps:
S51: take Least Square method to carry out fitting a straight line to the coordinate data in cut-off rule segment data, forms straight-line equation y=k*x+b; Y represents ordinate, k straight slope, and x represents horizontal ordinate, and b represents intercept on ordinate.
S52: select first coordinate data in line segment coordinate data and last coordinate data, and respectively vertical line is done to straight-line equation y=k*x+b, origin coordinates and the end coordinate of this straight-line equation is obtained by the intersection point calculating described vertical line and described straight-line equation;
S53: convert straight-line equation to polar equation;
S54: the differential seat angle absolute value of the adjacent segments in computed segmentation segment data, and the range difference absolute value of adjacent segments in cut-off rule segment data;
S55: judge whether differential seat angle absolute value is less than angle fixed threshold, and whether range difference absolute value is less than distance fixed threshold; If so, then two adjacent segment line segments are carried out merging and form renewal line segment, and replace the line segment before merging to upgrade line segment; If not, then next step is entered;
S56: return step S51 and be cycled to repeat, until all cut-off rule segment datas all calculate complete.
S6: until all cut-off rule segment datas merge complete and export merging segment data, described merging segment data is as the extraction segment data of laser radar range image.
Embodiment 2
As shown in Figure 2, the region segmentation process flow diagram that provides for the embodiment of the present invention of Fig. 2; Region segmentation method provided by the invention, concrete steps are as follows:
To the middle position value filtering method of regular length be adopted obtained laser radar data (p 1, p 2p i) carry out filtering process after, carry out region segmentation to this laser radar data, image of namely adjusting the distance adopts adaptive threshold value to carry out Region dividing, mainly comprises following 3 parts:
1) according to the sweep limit of laser radar and angular resolution by range data (p 1, p 2p i) convert polar coordinates (θ to 1, p 1), (θ 2, p 2) ... (θ i, p i); Within a scan period, obtaining environmental information from 2D laser radar is one group of range data (p 1, p 2p i), exceeding the range data that laser radar range obtains is 0; According to sweep limit and the angular resolution of this radar, can range data be converted to polar coordinates (θ i, p i) form, material is thus formed laser radar and original range image obtained to environment sensing.
2) according to polar coordinates (θ 1, p 1), (θ 2, p 2) ... (θ i, p i), adopt adaptive threshold, Data Segmentation is become region (D 1, D 2d m), wherein, D m(θ i, p i) set, refer to be partitioned into region.Partition principle has two, first principle be according to range data before and after change compared with first threshold theta1, second principle is that number containing 0 data between data is compared with Second Threshold theta2; Choosing of theta1 is dynamic, if current i=k, then first threshold theta1 value is p kwith p k-1difference add p k-1with p k-2difference;
I.e. i=flag, theta1=p i-p i-1+ p i-1-p i-2, theta2=α, α are the fixed error of laser radar; Get fixed threshold 10; (θ i, p i) refer to original angle-range coordinate;
First region segmentation of the present embodiment is not iteration, and after meeting segmentation condition, repetitive cycling iterations, until cover all data;
3) to region (D 1, D 2d m) in interference region process, give up interference region.
Embodiment 3
As shown in Figure 3, the line segment that Fig. 3 provides for the embodiment of the present invention splits process flow diagram; Line segment dividing method provided by the invention, concrete steps are as follows:
Line segment segmentation is carried out to each area data D, by the linear feature of line segment, region D is divided into (L 1, L 2l n), mainly comprise following 3 parts:
1) coordinate conversion: to regional ensemble D imiddle polar coordinates (θ 1, p 1), (θ 2, p 2) ... (θ i, p i) convert rectangular coordinate (x to 1, y 1), (x 2, y 2) ... (x i, y i), wherein, X i=p i* Cos θ i, Y i=p i* Sin θ i.
2) choosing of threshold value is adaptive, if current i=k, then theta value is the spacing sum of current point and front and back point.
To regional ensemble (D 1, D 2d i) in each region D iagain split, if D idata in set are (x 1, y 1), (x 2, y 2) ... (x n, y n),
With first coordinate points (x 1, y 1) and last coordinate points (x n, y n) simulate straight line:
(y n-y 1) * x-(x n-x 1) * y+y 1* x n-y n* x 1=0, find other coordinate points (x in region 2, y 2) ... (x n-1, y n-1) to the point of this straight line ultimate range, calculate ultimate range; The point of this straight line ultimate range, if meet segmentation condition, namely ultimate range is greater than threshold value, then with this point, this region segmentation is become two line segments.This region segmentation is become L 1, L 2, wherein L irepresenting a line segment, is (x n, y n) set, then upgrade whole region D i.
The process of an iteration to the segmentation in region, to two straight line L 1, L 2continue to adopt same dividing method.Until all linearity region (L 1, L 2l n) all meet line segment feature;
There is larger error in order to avoid common least square method (OLS) all to exist under error matching in situation at independent variable and dependent variable in the present embodiment, so adopt total least square method (TLS) to carry out matching to straight line.
Wherein choosing of threshold value calculates according to adaptive approach:
t h e t a = ( x k - x k - 1 ) 2 + ( y k - y k - 1 ) 2 + ( x k + 1 - x k ) 2 + ( y k + 1 - y k ) 2 , If the fixed error of laser radar is a, if theta<a, then theta gets fixed threshold a.After segmentation terminates, form (L 1, L 2l n).
The present embodiment is by the universal model y=k of line segment i* x+b iconvert (θ to i, p i) model form.Cross the vertical line that initial point makes line correspondence, wherein p icorresponding initial point to the distance of straight line, θ ithe angle that corresponding intersection point and X-axis are formed.Then (θ is utilized i, p i) model adopt threshold value go merge line segment, prevent over-segmentation.
Adopt and overcome when line segment and X-axis are close to time vertical in this way, k ibe worth excessive, cause the shortcoming that line segment pooled error is larger.Therefore, based on this, the line-fitting algorithm that the present embodiment provides both ensure that and the rapidity of traditional algorithm had improve adaptability and accuracy to a certain extent again.
3) to (L 1, L 2l n) in interference line segment process, give up interference line segment; Namely to line segment aggregate (L 1, L 2l n) in L iif, (x contained in L n, y n) number of coordinates be less than 3, be considered as disturb line segment, given up.
Wherein, D mreferring to region, is the set of line segment, L ireferring to the line segment inside region, is the set of coordinate (x, y); D m.length refer to the line segment number contained in this region, lower of starting condition is containing a L 1; (x k, y k) give directions (x 1, y 1), (x n, y n) connecting the point of straight line maximum distance, Max_Dis refers to the maximum distance of this point to straight line, and theta refers to segmentation adaptive threshold, if meet segmentation condition, is then cut-point with k, by L ibe divided into L tempt1, L tempt2, wherein use L tempt1replace L i, D m.Insert (i, L tempt2) refer to L tempt2be inserted into region D mmiddle L iafter.
Embodiment 4
As shown in Figure 4, the line segment that Fig. 4 provides for the embodiment of the present invention merges process flow diagram, line segment merging method provided by the invention, and concrete steps are as follows:
To the line segment (L split 1, L 2l n) adjacent line segment merges according to certain feature, prevents over-segmentation, mainly comprise with 4 steps:
1) respectively to L iin the coordinate data (x that comprises 1, y 1), (x 2, y 2) ... (x k, y k) carry out fitting a straight line, because variable X, Y contains stochastic error, in order to avoid common least square method (OLS) matching under independent variable and dependent variable all exist error condition exists larger error, so the method for matching adopts Least Square method (TLS).
2) according to L after matching istraight-line equation y=k i* x+b i, L iwith in first data point (x 1, y 1) and last data point (x k, y k) respectively vertical line is done to straight-line equation, by determining origin coordinates and end coordinate on straight-line equation, finally determine line segment, intersection point is origin coordinates and the end coordinate of this line segment.
3) by L imodel y=k i* x+b iconvert (θ to i, p i) model, cross the vertical line that initial point makes line correspondence, wherein p iinitial point to the distance of straight line, θ ithe angle that corresponding intersection point and X-axis are formed ,-180 ° of < θ i< 180 °.
4) according to (L in the D of region 1, L 2l n) straight line model (θ 1, p 1), (θ 2, p 2) ... (θ n, p n), line segment adjacent in regional ensemble is merged, when the absolute value of the difference of adjacent segments angle is less than fixed threshold and the absolute value of the difference of distance is less than fixed threshold, then two sections of line segments is merged, and update area D.The method of same employing iteration, until all adjacent line segments are all satisfied.
Wherein, D in Fig. 4 mreferring to a region, is the set of line segment, L irefer to the line segment inside region, D m.length the line segment number contained in this region is referred to.(θ i, p i) be straight line L istraight line model, ang_theta is angle threshold, and dis_theta is distance threshold, threshold value when it is merging.If meet merging condition, then by L iwith L i-1be merged into L tempt, replace L i-1, remove L simultaneously i, D m.remove (L i) refer to remove D ml in region i.
What finally illustrate is, above embodiment is only in order to illustrate technical scheme of the present invention and unrestricted, although by referring to the preferred embodiments of the present invention, invention has been described, but those of ordinary skill in the art is to be understood that, various change can be made to it in the form and details, and not depart from the spirit and scope that the present invention limits.

Claims (5)

1. a 2D laser radar range image middle conductor feature Accurate Curve-fitting method, is characterized in that: comprise the following steps:
S1: obtain laser radar range image data and carry out data prediction;
S2: be Polar Coordinate Model data by range image data transformations after data prediction;
S3: to Polar Coordinate Model data acquisition adaptive threshold region segmentation forming region data;
S4: line segment segmentation is carried out to each area data and forms cut-off rule segment data;
S5: line segment merging is carried out to the segment data in cut-off rule segment data;
S6: until all cut-off rule segment datas merge complete and export merging segment data, described merging segment data is as the extraction segment data of laser radar range image.
2. 2D laser radar range image middle conductor feature Accurate Curve-fitting method according to claim 1, is characterized in that: in described S1, data prediction filters random disturbance data in range image data by choosing middle position value filtering method.
3. 2D laser radar range image middle conductor feature Accurate Curve-fitting method according to claim 1, is characterized in that: in described S3, adaptive threshold region segmentation concrete steps are as follows:
S31: choose current data from Polar Coordinate Model data;
S32: judge whether the range data in current data is greater than zero, if so, then using current data sequence number as number label flag;
S33: if not, then returning step S31, to choose next Polar Coordinate Model data be current data;
S34: the front and back changing value calculating current data and upper data; Calculate the zero data number contained between current data and upper non-vanishing Polar Coordinate Model data;
S35: judge whether front and back changing value is greater than first threshold theta1, or whether zero data number is greater than Second Threshold theta2, if so, then current data is divided into m+1 cut zone data, and returning step S31, to choose next Polar Coordinate Model data be current data;
S36: if not, be then divided into m cut zone data by current data; And returning step S31, to choose next Polar Coordinate Model data be current data;
S37: to the last Polar Coordinate Model data;
S38: the number calculating contained data in each cut zone, is less than being considered as interference region and giving up of predetermined threshold value by number.
4. 2D laser radar range image middle conductor feature Accurate Curve-fitting method according to claim 1, is characterized in that: the line segment segmentation in described S4 comprises the following steps:
S41: convert the polar data of the area data in each cut zone to rectangular coordinate data;
S42: select two area datas in cut zone and go out straight line according to its rectangular coordinate data fitting;
S43: calculate the point of this area data other data interior to this straight line ultimate range, and calculate ultimate range;
These data if so, are then arranged at the first straight-line data region by S44: judge whether ultimate range is greater than adaptive threshold;
If not, then these data are arranged at the second straight-line data region;
Wherein, described adaptive threshold according to ultimate range according to following formulae discovery:
t h e t a = ( x k - x k - 1 ) 2 + ( y k - y k - 1 ) 2 + ( x k + 1 - x k ) 2 + ( y k + 1 - y k ) 2 , Wherein, (x k-1, y k-1) represent kth-1 rectangular coordinate data, (x k, y k) represent a kth rectangular coordinate data, (x k+1, y k+1) represent kth+1 rectangular coordinate data;
S45: be cycled to repeat and the first straight-line data region and the second straight-line data region are split, until all Data Segmentations are complete;
S46: the number calculating contained segmentation straight line in each cut zone, is less than being considered as interference region and giving up of predetermined threshold value by number.
5. 2D laser radar range image middle conductor feature Accurate Curve-fitting method according to claim 1, is characterized in that: described S5 middle conductor merging specifically comprises the following steps:
S51: take Least Square method to carry out fitting a straight line to the coordinate data in cut-off rule segment data, forms straight-line equation y=k*x+b;
S52: select first coordinate data in line segment coordinate data and last coordinate data, and respectively vertical line is done to straight-line equation y=k*x+b, origin coordinates and the end coordinate of this straight-line equation is obtained by the intersection point calculating described vertical line and described straight-line equation;
S53: convert straight-line equation to polar equation;
S54: the differential seat angle absolute value of the adjacent segments in computed segmentation segment data, and the range difference absolute value of adjacent segments in cut-off rule segment data;
S55: judge whether differential seat angle absolute value is less than angle fixed threshold, and whether range difference absolute value is less than distance fixed threshold;
If so, then two adjacent segment line segments are carried out merging and form renewal line segment, and replace the line segment before merging to upgrade line segment;
If not, then next step is entered;
S56: return step S51 and be cycled to repeat, until all cut-off rule segment datas all calculate complete.
CN201510511455.4A 2015-08-19 2015-08-19 A kind of 2D laser radars range image middle conductor feature Accurate Curve-fitting method Active CN105137412B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510511455.4A CN105137412B (en) 2015-08-19 2015-08-19 A kind of 2D laser radars range image middle conductor feature Accurate Curve-fitting method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510511455.4A CN105137412B (en) 2015-08-19 2015-08-19 A kind of 2D laser radars range image middle conductor feature Accurate Curve-fitting method

Publications (2)

Publication Number Publication Date
CN105137412A true CN105137412A (en) 2015-12-09
CN105137412B CN105137412B (en) 2017-10-20

Family

ID=54722818

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510511455.4A Active CN105137412B (en) 2015-08-19 2015-08-19 A kind of 2D laser radars range image middle conductor feature Accurate Curve-fitting method

Country Status (1)

Country Link
CN (1) CN105137412B (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105987697A (en) * 2016-04-26 2016-10-05 重庆大学 Right-angle bend Mecanum wheel type AGV (Automatic Guided Vehicle) navigation and positioning method and system
CN108805896A (en) * 2018-05-07 2018-11-13 南京理工大学 A kind of range image segmentation method applied to urban environment
CN109443370A (en) * 2018-09-13 2019-03-08 中通国脉物联科技南京有限公司 A method of deviate detection track
CN110146910A (en) * 2019-05-15 2019-08-20 重庆大学 A kind of localization method and device merged based on GPS with laser radar data
CN110399892A (en) * 2018-04-24 2019-11-01 北京京东尚科信息技术有限公司 Environmental characteristic extracting method and device
CN110412530A (en) * 2018-04-27 2019-11-05 深圳市优必选科技有限公司 Method and device for identifying charging pile and robot
CN110632607A (en) * 2019-09-16 2019-12-31 广州小鹏汽车科技有限公司 Object boundary determining method and system based on ultrasonic data and vehicle
CN113093219A (en) * 2021-02-20 2021-07-09 意欧斯物流科技(上海)有限公司 2D laser radar data line segment extraction method based on region growing algorithm

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1808711A2 (en) * 2006-01-16 2007-07-18 Omron Corporation Object detector
CN102930280A (en) * 2012-10-05 2013-02-13 中国电子科技集团公司第十研究所 Method for identifying overhead high-voltage wire automatically from infrared image
CN103760569A (en) * 2013-12-31 2014-04-30 西安交通大学 Drivable region detection method based on laser radar
CN103777192A (en) * 2012-10-24 2014-05-07 中国人民解放军第二炮兵工程学院 Linear feature extraction method based on laser sensor
CN104008387A (en) * 2014-05-19 2014-08-27 山东科技大学 Lane line detection method based on feature point piecewise linear fitting
CN104200212A (en) * 2014-06-25 2014-12-10 西安煤航信息产业有限公司 Building outer boundary line extraction method based on onboard LiDAR (Light Detection and Ranging) data
CN104463871A (en) * 2014-12-10 2015-03-25 武汉大学 Streetscape facet extraction and optimization method based on vehicle-mounted LiDAR point cloud data

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1808711A2 (en) * 2006-01-16 2007-07-18 Omron Corporation Object detector
CN102930280A (en) * 2012-10-05 2013-02-13 中国电子科技集团公司第十研究所 Method for identifying overhead high-voltage wire automatically from infrared image
CN103777192A (en) * 2012-10-24 2014-05-07 中国人民解放军第二炮兵工程学院 Linear feature extraction method based on laser sensor
CN103760569A (en) * 2013-12-31 2014-04-30 西安交通大学 Drivable region detection method based on laser radar
CN104008387A (en) * 2014-05-19 2014-08-27 山东科技大学 Lane line detection method based on feature point piecewise linear fitting
CN104200212A (en) * 2014-06-25 2014-12-10 西安煤航信息产业有限公司 Building outer boundary line extraction method based on onboard LiDAR (Light Detection and Ranging) data
CN104463871A (en) * 2014-12-10 2015-03-25 武汉大学 Streetscape facet extraction and optimization method based on vehicle-mounted LiDAR point cloud data

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
满增光 等: "基于自适应阈值的距离图像线段特征提取", 《深圳大学学报理工版》 *

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105987697B (en) * 2016-04-26 2019-01-29 重庆大学 The wheeled AGV navigation locating method of Mecanum and system under a kind of quarter bend
CN105987697A (en) * 2016-04-26 2016-10-05 重庆大学 Right-angle bend Mecanum wheel type AGV (Automatic Guided Vehicle) navigation and positioning method and system
CN110399892A (en) * 2018-04-24 2019-11-01 北京京东尚科信息技术有限公司 Environmental characteristic extracting method and device
CN110399892B (en) * 2018-04-24 2022-12-02 北京京东尚科信息技术有限公司 Environmental feature extraction method and device
CN110412530B (en) * 2018-04-27 2021-09-17 深圳市优必选科技有限公司 Method and device for identifying charging pile and robot
CN110412530A (en) * 2018-04-27 2019-11-05 深圳市优必选科技有限公司 Method and device for identifying charging pile and robot
CN108805896B (en) * 2018-05-07 2022-02-18 南京理工大学 Distance image segmentation method applied to urban environment
CN108805896A (en) * 2018-05-07 2018-11-13 南京理工大学 A kind of range image segmentation method applied to urban environment
CN109443370A (en) * 2018-09-13 2019-03-08 中通国脉物联科技南京有限公司 A method of deviate detection track
CN110146910A (en) * 2019-05-15 2019-08-20 重庆大学 A kind of localization method and device merged based on GPS with laser radar data
CN110146910B (en) * 2019-05-15 2023-02-28 重庆大学 Positioning method and device based on data fusion of GPS and laser radar
CN110632607A (en) * 2019-09-16 2019-12-31 广州小鹏汽车科技有限公司 Object boundary determining method and system based on ultrasonic data and vehicle
CN113093219A (en) * 2021-02-20 2021-07-09 意欧斯物流科技(上海)有限公司 2D laser radar data line segment extraction method based on region growing algorithm

Also Published As

Publication number Publication date
CN105137412B (en) 2017-10-20

Similar Documents

Publication Publication Date Title
CN105137412A (en) Accurate fitting method of line segment features in 2D laser radar distance image
KR102143108B1 (en) Lane recognition modeling method, device, storage medium and device, and recognition method, device, storage medium and device
CN110738121A (en) front vehicle detection method and detection system
CN101916373B (en) Road semiautomatic extraction method based on wavelet detection and ridge line tracking
Gikas et al. A novel geodetic engineering method for accurate and automated road/railway centerline geometry extraction based on the bearing diagram and fractal behavior
CN112763995B (en) Radar calibration method and device, electronic equipment and road side equipment
CN102136155A (en) Object elevation vectorization method and system based on three dimensional laser scanning
CN108279436A (en) Method and device for picking up seismic wave first arrivals
CN103017655B (en) Method and system for extracting floor area of multi-floor building
CN107730582B (en) Ocean wave three-dimensional display method based on ocean remote sensing data
CN111007531A (en) Road edge detection method based on laser point cloud data
CN114565616B (en) Unstructured road state parameter estimation method and system
CN114547866B (en) Prefabricated part intelligent detection method based on BIM-unmanned aerial vehicle-mechanical dog
CN112379393B (en) Train collision early warning method and device
CN109919955A (en) The tunnel axis of ground formula laser radar point cloud extracts and dividing method
CN108305277A (en) A kind of heterologous image matching method based on straightway
CN114332291A (en) Oblique photography model building outer contour rule extraction method
CN113325389A (en) Unmanned vehicle laser radar positioning method, system and storage medium
CN113970734A (en) Method, device and equipment for removing snowing noise of roadside multiline laser radar
CN114283343B (en) Map updating method, training method and device based on remote sensing satellite image
JP6322564B2 (en) Point cloud analysis processing apparatus, method, and program
CN116222577A (en) Closed loop detection method, training method, system, electronic equipment and storage medium
CN103901467A (en) Method for tracking positions of three-dimensional seismic data
Lu et al. A lightweight real-time 3D LiDAR SLAM for autonomous vehicles in large-scale urban environment
CN114488026A (en) Underground parking garage passable space detection method based on 4D millimeter wave radar

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right

Effective date of registration: 20201222

Address after: 402460 station No.14, no.6, 10th floor, innovation and development center, No.19 Lingfang Avenue, Changzhou street, Rongchang District, Chongqing

Patentee after: Chongqing kezhiyuan Technology Co.,Ltd.

Address before: 400044 No. 174 Sha Jie street, Shapingba District, Chongqing

Patentee before: Chongqing University

TR01 Transfer of patent right