CN202134079U - Unmanned vehicle lane marker line identification and alarm device - Google Patents
Unmanned vehicle lane marker line identification and alarm device Download PDFInfo
- Publication number
- CN202134079U CN202134079U CN201120203640U CN201120203640U CN202134079U CN 202134079 U CN202134079 U CN 202134079U CN 201120203640 U CN201120203640 U CN 201120203640U CN 201120203640 U CN201120203640 U CN 201120203640U CN 202134079 U CN202134079 U CN 202134079U
- Authority
- CN
- China
- Prior art keywords
- unmanned vehicle
- image
- pick
- head
- automatic driving
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The utility model discloses an unmanned vehicle lane marker line identification and alarm device, comprising an unmanned vehicle, wherein a pick-up head is arranged on a longitudinal center line in the front of the unmanned vehicle, the pick-up head transmits acquired images to an image processing module, the image processing module is connected with a deviation determining analysis module, and the deviation determining analysis module is connected with an alarm. Compared with an unmanned vehicle adopting radar or a laser sensor, the marker line identification and alarm device adopts the pick-up head as a main means for the unmanned vehicle to sensor environment, thereby fully utilizing a computer and a loudspeaker of the current unmanned vehicle. The device is low in investment, ad simple and easy to operate. Only the pick-up head needs to be installed, and the pick-up head can better simulate the visual field of a driver, thereby providing convenience for the unmanned vehicle to simulate the driving of a people.
Description
Technical field
The utility model relates to the automatic driving vehicle technical field, specifically, relates to a kind of automatic driving vehicle traffic lane line identification and warning device.
Background technology
Pilotless automobile integrates numerous technology such as automatic control, architecture, artificial intelligence, vision calculating; It is the product of computer science, pattern-recognition and intelligent control technology high development; Also be an important symbol weighing national scientific research strength and industrial level, have broad application prospects in national defence and national economy field.Automatic driving vehicle is studied by countries in the world as the developing direction of following automobile widely.Since the seventies in 20th century, developed countries such as the U.S., Britain, Germany begin to carry out the research of pilotless automobile, are all obtaining breakthrough progress aspect feasibility and the practicability at present.China carries out the research of pilotless automobile since the eighties in 20th century, has also obtained certain achievement at present.Compare the automatic driving vehicle that adopts radar and laser sensor; Adopt the main means of camera as the pilotless automobile environment sensing; Have and drop into advantages such as low, simple and easy to do; And the image of camera collection more can drive simulating person the visual field, more helpful for automatic driving vehicle anthropomorphic dummy's driving.
Summary of the invention
Automotive run-off-road warning system complex structure to the prior art existence; The shortcoming that cost is high; The purpose of the utility model is; A kind of warning device of automatic driving vehicle run-off-road markings is provided, and this device identifies the curb of road both sides through Flame Image Process, thereby controlled variable is provided for the control that turns to of automatic driving vehicle.
In order to realize above-mentioned task, the utility model is taked following technical solution:
A kind of automatic driving vehicle traffic lane line identification and warning device; Comprise automatic driving vehicle; It is characterized in that on automatic driving vehicle front end longitudinal centre line, camera is installed, camera is transferred to image processing module with the image that collects; Image processing module with depart from the discriminatory analysis module and be connected, depart from the discriminatory analysis module being connected with alarm.
Other characteristics of the utility model are that the image processing module that is adopted is used for the image gray processing that collects, Filtering Processing, histogram equalization, rim detection and binary conversion treatment, expansion and corrosion treatment and Hough conversion, extracts traffic lane line.
Described warning device is a vehicle-mounted loudspeaker.
The automatic driving vehicle that adopts radar and laser sensor is compared in the automatic driving vehicle traffic lane line identification of the utility model with warning device; Adopt the main means of camera as the pilotless automobile environment sensing; Made full use of existing automatic driving vehicle on computing machine and loudspeaker; Have and drop into advantages such as low, simple and easy to do; Only needing increases shooting, and the visual field that the image of camera collection more can drive simulating person, more helpful for automatic driving vehicle anthropomorphic dummy's driving.
Description of drawings
Fig. 1 is the structural representation of the utility model;
Fig. 2 is the utility model image processing module principle flow chart;
Fig. 3 is the right-angle triangle schematic diagram;
Below in conjunction with accompanying drawing the utility model is done further to specify.
Embodiment
As shown in Figure 1; Identification of automatic driving vehicle traffic lane line and warning device; Comprise automatic driving vehicle, on automatic driving vehicle front end longitudinal centre line, camera is installed, camera is transferred to image processing module with the image that collects; Image processing module with depart from the discriminatory analysis module and be connected, depart from the discriminatory analysis module being connected with alarm.
Camera is after road pavement is taken continuously; The image that collects is transferred to image processing module; Image processing module is based on the matlab software platform; After the image that collects carried out gray processing, Filtering Processing, histogram equalization, rim detection and binary conversion treatment, expansion and corrosion treatment, Hough conversion, traffic lane line is extracted.At last markings are departed from discriminatory analysis, the terminal connects alarm.
In the present embodiment, alarm is a vehicle-mounted loudspeaker.
As shown in Figure 2, image processing module is received and is carried out following image processing step behind the image that collects:
1, the gray scale of image is handled:
The bright-dark degree that each pixel had in the image is identified by gray-scale value (gray level).Generally the gray-scale value with white is defined as 255, and the black gray value defined is 0, and is divided into 256 grades equably by black shading value in vain.So earlier the image that collects being carried out gray scale handles.
2, the Filtering Processing of image:
Any-a undressed original image, all exist noise to a certain degree.Noise penalty picture quality, make image blurringly, even flood characteristic, bring difficulty to analysis.The work of removal of images noise is referred to as image smoothing or filtering.Level and smooth purpose has two: improve picture quality and extract characteristics of objects out.In the images acquired process; Except receiving noise effect, also receive the influence of quantization error; It makes the image border become unintelligible, and image processing process adopts the bidimensional convolution algorithm that gray-scale map is carried out smothing filtering, and this computing is the process of weighted sum; Each pixel in the image-region that uses multiplies each other with each corresponding element of convolution kernel respectively, and all sum of products are as the new value of regional center pixel.The concrete formula of bidimensional convolution algorithm is:
With
around its initial point Rotate 180 degree; Translation initial point then; U axially goes up translation x, and v axially goes up translation y.Again with the output that integration can get a some place of multiplying each other of two functions.The weighted sum of this computing utilization input neighborhood of pixels pixel substitutes the input pixel, can make the border of gray-scale map become more level and smooth.
3, rim detection:
Because traffic lane line and road surface background parts have more intense contrast, so edge detection algorithm capable of using detects the traffic lane line edge.In detecting on the edge of, a kind of template commonly used is the Sobel operator, because the Sobel operator is the form of filter operator, can utilize the fast convolution function, and is effectively simple, therefore is widely used.Simultaneously, compare with the Prewitt operator, the Sobel operator has been done weighting for the influence of pixel location, so better effects if.The Sobel operator has two, and one is the detection level edge, and promptly [1-2 1; 000; 12 1]; Another is the flat edge of detection of vertical, promptly [1 01;-202;-10 1].
4, binary conversion treatment:
Binaryzation is claimed the gray scale graduation again, and every needs are done the image of route identification, all can utilize this mode.The basic process of binaryzation is following: earlier original image is done middle LPF, carry out the image pre-service, reduce or remove noise, confirm best threshold values with algorithm again, every grey scale pixel value is set as 255 greater than this threshold values, is set as 0 less than this threshold values.Image after handling like this just has only black-and-white two color, thereby tonal range is divided into two types of target and backgrounds, realizes image binaryzation.Binaryzation not only can strengthen road boundary, and for the real-time that reduces amount of image information and enhancement algorithms better effect is arranged after the binaryzation.
Generally can be divided into two types of overall threshold values algorithm and local threshold values algorithms to the threshold values choosing method of image pixel.Overall situation threshold values algorithm is to choose a fixing threshold values with image binaryzation according to entire image.Overall threshold values algorithm commonly used has big law (Otsu method) etc.Overall situation threshold values algorithm is fairly simple, realizes than is easier to, and the grey level histogram that is applicable to image has significantly bimodal, and the lowest point corresponding gray of optional grey level histogram is best threshold values at this moment.Local threshold values algorithm is to be the plurality of sub image with image division, in conjunction with the pixel of current investigation and the gray-scale value of its neighborhood territory pixel point, and a threshold values of confirming to investigate.
Select the Otsu method here for use, promptly maximum variance between clusters carries out binaryzation.This algorithm is based on the statistical property of view picture figure, realizes choosing automatically of threshold values.Its basic idea is with the gray-scale value of a certain supposition the gray scale of image to be divided into two types, and when two groups inter-class variances were maximum, this gray-scale value was exactly the best threshold values of image binaryzation.Let images have
gray value range at
, the gray scale value selected within the range
, the image is divided into two groups
and
,
contains the pixel gray values
,
gray values
, with
represents the total number of image pixels,
represent gray value
is the number of pixels, then each grayscale value
probability is
;
and
class probability and mean: probability
; mean
, class variance
, the best threshold
is to make the maximum between-class variance
values, ie
.
5, expansion corrosion treatment:
Expansion is that an operator handling of morphology is for bianry image; Expansion means; Moving on image with structural element, is overlapping just passable as long as point in a point and the image is arranged in the structural element, and the result of expansion is the inside in the structural element central point zone of streaking.So edge of image has just been enlarged.For gray level image; Expansion means; In the scope of structural element, ask for corresponding element in structural element and the image with, see which and maximum; Then this and just as the value of image center, certainly this central point be meant with that corresponding picture position of structural element center on.Continue the slide construction element then.Corrosion can be regarded the dual operations of expansion as.
Image also exists some noises after through filtering and rim detection, receive these interference of noise in order to reduce traffic lane line, and image processing module adopts the processing procedure of first expansion post-etching to eliminate residual noise.Utilize expansion algorithm can make certain neighborhood of pixels as long as there is a white pixel, this pixel will become white from black so, and remaining remains unchanged; In contrast, erosion algorithm makes certain neighborhood of pixels as long as there is a black picture element, and this pixel will become black from white so, and remaining remains unchanged.Because salt-pepper noise is exactly white (deceiving) point on black (in vain) image, so utilize above-mentioned principle, can well eliminate through the expansion corrosion process and leave over noise.
6, lane line extracts
Through having comprised a lot of spuious lines in the image after the image pre-service, so need discern and extract to lane line.This paper adopts the Hough conversion to detect traffic lane line.The thought of Hough conversion is: in a coordinates of original image coordinates system point correspondence down the straight line in the parameter coordinate system; The straight line of same parameter coordinate system is a corresponding point under the original coordinate system; Appear under the original coordinate system then straight line have a few; Their slope is identical with intercept, all their corresponding same points under the parameter coordinate system.Like this with each spot projection under the original coordinate system to the parameter coordinate system down after, see whether convergence point is arranged under the parameter coordinate system, such convergence point just correspondence the straight line under the original coordinate system.The polar equation of straight line is:
Using the Hough transform to a point on the line
is converted into two-dimensional parameter space
on the spot.Then
changed into zone of dispersion, the quantity of the point that falls into this zone of adding up.After conversion is accomplished; The quantity that adds up zone how is just corresponding to a common ground on two-dimensional parameter space
, and
is exactly the Straight Line Fitting Parameters of image space.In coordinates of original image coordinates, just can extract traffic lane line through proper transformation again.Image processing module is followed the tracks of the traffic lane line that extracts through the Kalman wave filter again; The location status that dopes traffic lane line in the next frame image line parameter of going forward side by side calculates; And this parameter passed to the Hough conversion; So that on this parameter basis, carry out the detection of the traffic lane line of next frame image, reach the purpose that reduces Hough transformation calculations amount.
After detecting traffic lane line, can confirm the distance between vehicle and the markings, thereby open alarm according to following right-angle triangle principle.
As shown in Figure 3, vehicle body constitutes a right-angle triangle to distance, traffic lane line gusset profile and the traffic lane line three of traffic lane line, if the automatic driving vehicle cruising then should keep such shape in the zone always.Carrying out automatic driving vehicle when control, automatic driving vehicle should have a safe distance standard from graticule, and this right-angle triangle base this safe distance just, can be through confirming that this safe distance threshold values define.In case vehicle body, just starts alarm and reports to the police less than this threshold values to the distance of traffic lane line.
Claims (2)
1. an automatic driving vehicle traffic lane line is discerned and warning device; Comprise automatic driving vehicle; It is characterized in that on automatic driving vehicle front end longitudinal centre line, camera is installed, camera is transferred to image processing module with the image that collects; Image processing module with depart from the discriminatory analysis module and be connected, depart from the discriminatory analysis module being connected with alarm.
2. automatic driving vehicle traffic lane line identification as claimed in claim 1 and warning device is characterized in that described vehicle-mounted loudspeaker.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201120203640U CN202134079U (en) | 2011-06-16 | 2011-06-16 | Unmanned vehicle lane marker line identification and alarm device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201120203640U CN202134079U (en) | 2011-06-16 | 2011-06-16 | Unmanned vehicle lane marker line identification and alarm device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN202134079U true CN202134079U (en) | 2012-02-01 |
Family
ID=45522842
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201120203640U Expired - Fee Related CN202134079U (en) | 2011-06-16 | 2011-06-16 | Unmanned vehicle lane marker line identification and alarm device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN202134079U (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529404A (en) * | 2016-09-30 | 2017-03-22 | 张家港长安大学汽车工程研究院 | Imaging principle-based recognition method for pilotless automobile to recognize road marker line |
| CN106886217A (en) * | 2017-02-24 | 2017-06-23 | 安科智慧城市技术（中国）有限公司 | Automatic navigation control method and apparatus |
| CN107389084A (en) * | 2017-06-09 | 2017-11-24 | 深圳市速腾聚创科技有限公司 | Planning driving path planing method and storage medium |
| CN107833492A (en) * | 2017-11-24 | 2018-03-23 | 南京视莱尔汽车电子有限公司 | A kind of zebra stripes real-time monitoring system based on intelligent driving |
| CN108416320A (en) * | 2018-03-23 | 2018-08-17 | 京东方科技集团股份有限公司 | Inspection device, the control method of inspection device and control device |
| CN109398356A (en) * | 2018-11-23 | 2019-03-01 | 奇瑞汽车股份有限公司 | Lane Keeping System and method |
| CN111862226A (en) * | 2019-04-30 | 2020-10-30 | 百度（美国）有限责任公司 | Hardware design for camera calibration and image pre-processing in a vehicle |
| CN113028912A (en) * | 2021-04-21 | 2021-06-25 | 湘潭大学 | Bullet primer priming charge filling amount detection method based on 3D vision |
-
2011
- 2011-06-16 CN CN201120203640U patent/CN202134079U/en not_active Expired - Fee Related
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529404A (en) * | 2016-09-30 | 2017-03-22 | 张家港长安大学汽车工程研究院 | Imaging principle-based recognition method for pilotless automobile to recognize road marker line |
| CN106886217A (en) * | 2017-02-24 | 2017-06-23 | 安科智慧城市技术（中国）有限公司 | Automatic navigation control method and apparatus |
| CN107389084A (en) * | 2017-06-09 | 2017-11-24 | 深圳市速腾聚创科技有限公司 | Planning driving path planing method and storage medium |
| CN107389084B (en) * | 2017-06-09 | 2020-06-05 | 深圳市速腾聚创科技有限公司 | Driving path planning method and storage medium |
| CN107833492A (en) * | 2017-11-24 | 2018-03-23 | 南京视莱尔汽车电子有限公司 | A kind of zebra stripes real-time monitoring system based on intelligent driving |
| CN108416320A (en) * | 2018-03-23 | 2018-08-17 | 京东方科技集团股份有限公司 | Inspection device, the control method of inspection device and control device |
| US10580124B2 (en) * | 2018-03-23 | 2020-03-03 | Boe Technology Group Co., Ltd. | Inspection device, control method and control apparatus for the same |
| CN109398356A (en) * | 2018-11-23 | 2019-03-01 | 奇瑞汽车股份有限公司 | Lane Keeping System and method |
| CN111862226A (en) * | 2019-04-30 | 2020-10-30 | 百度（美国）有限责任公司 | Hardware design for camera calibration and image pre-processing in a vehicle |
| CN111862226B (en) * | 2019-04-30 | 2024-01-16 | 百度（美国）有限责任公司 | Hardware design for camera calibration and image preprocessing in a vehicle |
| CN113028912A (en) * | 2021-04-21 | 2021-06-25 | 湘潭大学 | Bullet primer priming charge filling amount detection method based on 3D vision |
| CN113028912B (en) * | 2021-04-21 | 2022-09-20 | 湘潭大学 | Bullet primer priming charge filling amount detection method based on 3D vision |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN202134079U (en) | Unmanned vehicle lane marker line identification and alarm device | |
| CN102682292B (en) | Method based on monocular vision for detecting and roughly positioning edge of road | |
| Yan et al. | A method of lane edge detection based on Canny algorithm | |
| CN103714538B (en) | road edge detection method, device and vehicle | |
| CN105678285B (en) | A kind of adaptive road birds-eye view transform method and road track detection method | |
| Li et al. | Nighttime lane markings recognition based on Canny detection and Hough transform | |
| CN105488454A (en) | Monocular vision based front vehicle detection and ranging method | |
| CN102938057B (en) | A kind of method for eliminating vehicle shadow and device | |
| CN102354457B (en) | General Hough transformation-based method for detecting position of traffic signal lamp | |
| CN104392212A (en) | Method for detecting road information and identifying forward vehicles based on vision | |
| Wang et al. | A vision-based road edge detection algorithm | |
| CN104657735A (en) | Lane line detection method and system, as well as lane departure early warning method and system | |
| CN105260701A (en) | Front vehicle detection method applied to complex scene | |
| CN106802144A (en) | A kind of vehicle distance measurement method based on monocular vision and car plate | |
| CN106887004A (en) | A kind of method for detecting lane lines based on Block- matching | |
| CN105426868A (en) | Lane detection method based on adaptive region of interest | |
| KR20110001427A (en) | High speed road lane detection method based on extraction of roi-lb | |
| CN102982304A (en) | Method and system used for detecting vehicle positions by means of polarized images | |
| CN105678287A (en) | Ridge-measure-based lane line detection method | |
| CN105303157A (en) | Algorithm to extend detecting range for AVM stop line detection | |
| CN108198417A (en) | A kind of road cruising inspection system based on unmanned plane | |
| CN108304749A (en) | The recognition methods of road speed line, device and vehicle | |
| Li et al. | A lane marking detection and tracking algorithm based on sub-regions | |
| Li et al. | Judgment and optimization of video image recognition in obstacle detection in intelligent vehicle | |
| Chen | Road vehicle recognition algorithm in safety assistant driving based on artificial intelligence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120201 Termination date: 20120616 |