CN107133597A - A kind of front vehicles detection method in the daytime - Google Patents
A kind of front vehicles detection method in the daytime Download PDFInfo
- Publication number
- CN107133597A CN107133597A CN201710327872.2A CN201710327872A CN107133597A CN 107133597 A CN107133597 A CN 107133597A CN 201710327872 A CN201710327872 A CN 201710327872A CN 107133597 A CN107133597 A CN 107133597A
- Authority
- CN
- China
- Prior art keywords
- sample
- training
- fenton oxidation
- daytime
- vehicle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F18/2148—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the process organisation or structure, e.g. boosting cascade
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Abstract
The invention discloses a kind of front vehicles detection method in the daytime, comprise the following steps：S1：The video image of vehicle front is obtained by camera, video image is pre-processed；S2：Off-line training：Like-Fenton Oxidation is extracted by integrogram to the substantial amounts of positive and negative samples of vehicle in the daytime picture, choosing effective Like-Fenton Oxidation based on Adaboost algorithm is trained, and obtains strong classifier；S3：ONLINE RECOGNITION：Like-Fenton Oxidation is extracted to test sample, and inputs the feature into AdaBoost graders and carries out vehicle identification.The present invention is when extracting Like-Fenton Oxidation, and the use of integrogram is effectively improved training speed and detection speed, and quickly and efficiently vehicle can be identified.
Description
Technical field
Present invention relates particularly to a kind of front vehicles detection method in the daytime.
Background technology
Vehicle rear-end collision occupies significant proportion in traffic accident, is collided to be effectively prevented from vehicle, front vehicles
Detection technique turns into intelligence and the important research direction in safety driving assist system field.Front vehicles detecting system is by passing
Sensor provides front environmental information for driving vehicle.The sensing used at present mainly has machine vision, millimetre-wave radar, laser thunder
Sensor is waited up to infrared.Relative to other sensors, vision sensor has the big notable feature of obtained information quantity.Therefore,
Machine vision is still to realize the main sensors of front vehicles detection.
The vehicle identification method of view-based access control model can be generally divided into feature based, based on optical flow field, based on model and be based on
4 kinds of machine learning.The vehicle checking method of feature based is mainly carried out according to features such as the symmetry of vehicle, shade, edges
Vehicle detection, to obtain definite result, generally combines shade, symmetry and edge feature.Optical flow method is mainly by taking the photograph
The instantaneous velocity of camera motion, front obstacle motion or the two relative motion is realized, but this method changes to noise, light
Sensitivity, it is computationally intensive.Method based on model initially set up known vehicle accurately two dimension or threedimensional model, then with it is to be detected
Image is matched, and this method is undue to auto model to be relied on.Machine learning mainly utilizes the rule extracted in data or mould
Data are converted into information by formula, and Classification and Identification is carried out to data.Machine learning method based on SVM or neutral net is examined to vehicle
Survey hour operation quantity is big, time-consuming, and recognition performance needs further raising.
The content of the invention
The technical problem to be solved in the present invention is to provide a kind of front vehicles detection method in the daytime.
A kind of front vehicles detection method in the daytime, comprises the following steps：
S1：The video image of vehicle front is obtained by camera, video image is pre-processed, including at gray processing processing
Reason and normalized are 20 × 20 gray-scale maps；
S2：Off-line training：Like-Fenton Oxidation is extracted by integrogram to the substantial amounts of positive and negative samples of vehicle in the daytime picture, is based on
Adaboost algorithm is chosen effective Like-Fenton Oxidation and is trained, and obtains strong classifier；
S3：ONLINE RECOGNITION：Like-Fenton Oxidation is extracted to test sample, and inputs the feature into AdaBoost graders and carries out vehicle
Identification.
Further, the detailed process of off-line training is as follows：
1）Training sample selection：Training sample is divided into positive sample and negative sample, and positive sample is vehicle samples pictures, negative sample in the daytime
For other arbitrary sample pictures；From positive sample 1000, negative sample 3000 carries out gray processing processing, normalization to sample
It is processed as 20 × 20 gray-scale maps, composing training sample set；
2）Calculated using integrogram on image pixel in any rectangular area and, so as to try to achieve rectangle value；At point (x, y) place
Pixel summation from left to right is：
；
In formula：ForPoint integral image pixel and；For pointThe gray value at place；Borrow
Integral image is helped quickly to calculate Like-Fenton Oxidation value；
3）Based on Adaboost classifier trainings：
1. sample set：, whereinFor the training sample vector of input；For classification mark
Label,；0 and 1 represents negative sample and positive sample respectively, it is assumed that l is had in sample set
Individual positive sample, m negative sample, l+m=n, each sample has k Like-Fenton Oxidation value；
2. weight is initialized：
WhenWhen be negative sample,；WhenIt is positive sample when=1,；
3. for t=1,2 ..., T (T is frequency of training, determines the number of final Weak Classifier)：
Weight is normalized：；
For feature j, Weak Classifier is trained by given sample weights, and calculate its error relative to present weight：
In formula,For the value of Weak Classifier；For the characteristic value of j-th of feature；For threshold value；, table
Show classification direction；Select that there is error in epicycle trainingWeak ClassifierIt is added in strong classifier,；
Update the weight corresponding to each sample：
,
In formula,；If sampleCorrectly classified, then；Otherwise,；
4. final strong classifier is：
；In formula,。
The beneficial effects of the invention are as follows：
The present invention is when extracting Like-Fenton Oxidation, and the use of integrogram is effectively improved training speed and detection speed, can be with
Quickly and efficiently vehicle is identified.
Embodiment
The present invention is further elaborated for specific examples below, but not as a limitation of the invention.
A kind of front vehicles detection method in the daytime, comprises the following steps：
S1：The video image of vehicle front is obtained by camera, video image is pre-processed, including at gray processing processing
Reason and normalized are 20 × 20 gray-scale maps；
S2：Off-line training：Like-Fenton Oxidation is extracted by integrogram to the substantial amounts of positive and negative samples of vehicle in the daytime picture, is based on
Adaboost algorithm is chosen effective Like-Fenton Oxidation and is trained, and obtains strong classifier；
S3：ONLINE RECOGNITION：Like-Fenton Oxidation is extracted to test sample, and inputs the feature into AdaBoost graders and carries out vehicle
Identification.
The detailed process of off-line training is as follows：
1）Training sample selection：Training sample is divided into positive sample and negative sample, and positive sample is vehicle samples pictures, negative sample in the daytime
For other arbitrary sample pictures；From positive sample 1000, negative sample 3000 carries out gray processing processing, normalization to sample
It is processed as 20 × 20 gray-scale maps, composing training sample set；
2）Calculated using integrogram on image pixel in any rectangular area and, so as to try to achieve rectangle value；At point (x, y) place
Pixel summation from left to right is：
；
In formula：ForPoint integral image pixel and；For pointThe gray value at place；Borrow
Integral image is helped quickly to calculate Like-Fenton Oxidation value；
3）Based on Adaboost classifier trainings：
1. sample set：, whereinFor the training sample vector of input；For classification mark
Label,；0 and 1 represents negative sample and positive sample respectively, it is assumed that l is had in sample set
Individual positive sample, m negative sample, l+m=n, each sample has k Like-Fenton Oxidation value；
2. weight is initialized：
WhenWhen be negative sample,；WhenIt is positive sample when=1,；
3. for t=1,2 ..., T (T is frequency of training, determines the number of final Weak Classifier)：
Weight is normalized：；
For feature j, Weak Classifier is trained by given sample weights, and calculate its error relative to present weight：
In formula,For the value of Weak Classifier；For the characteristic value of j-th of feature；For threshold value；, table
Show classification direction；Select that there is error in epicycle trainingWeak ClassifierIt is added in strong classifier,；
Update the weight corresponding to each sample：
,
In formula,；If sampleCorrectly classified, then；Otherwise,；
4. final strong classifier is：
；In formula,。
Claims (2)
1. a kind of front vehicles detection method in the daytime, it is characterised in that comprise the following steps：
S1：The video image of vehicle front is obtained by camera, video image is pre-processed, including at gray processing processing
Reason and normalized are 20 × 20 gray-scale maps；
S2：Off-line training：Like-Fenton Oxidation is extracted by integrogram to the substantial amounts of positive and negative samples of vehicle in the daytime picture, is based on
Adaboost algorithm is chosen effective Like-Fenton Oxidation and is trained, and obtains strong classifier；
S3：ONLINE RECOGNITION：Like-Fenton Oxidation is extracted to test sample, and inputs the feature into AdaBoost graders and carries out vehicle
Identification.
2. front vehicles detection method in the daytime according to claim 1, it is characterised in that the detailed process of off-line training is such as
Under：
1）Training sample selection：Training sample is divided into positive sample and negative sample, and positive sample is vehicle samples pictures, negative sample in the daytime
For other arbitrary sample pictures；From positive sample 1000, negative sample 3000 carries out gray processing processing, normalization to sample
It is processed as 20 × 20 gray-scale maps, composing training sample set；
2）Calculated using integrogram on image pixel in any rectangular area and, so as to try to achieve rectangle value；At point (x, y) place
Pixel summation from left to right is：
；
In formula：ForPoint integral image pixel and；For pointThe gray value at place；By
Integral image can quickly calculate Like-Fenton Oxidation value；
3）Based on Adaboost classifier trainings：
1. sample set：, whereinFor the training sample vector of input；For classification mark
Label,；0 and 1 represents negative sample and positive sample respectively, it is assumed that l is had in sample set
Individual positive sample, m negative sample, l+m=n, each sample has k Like-Fenton Oxidation value；
2. weight is initialized：
WhenWhen be negative sample,；WhenIt is positive sample when=1,；
3. for t=1,2 ..., T (T is frequency of training, determines the number of final Weak Classifier)：
Weight is normalized：；
For feature j, Weak Classifier is trained by given sample weights, and calculate its error relative to present weight：
In formula,For the value of Weak Classifier；For the characteristic value of j-th of feature；For threshold value；, table
Show classification direction；Select that there is error in epicycle trainingWeak ClassifierIt is added in strong classifier,；
Update the weight corresponding to each sample：
,
In formula,；If sampleCorrectly classified, then；Otherwise,；
4. final strong classifier is：
；In formula,。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710327872.2A CN107133597A (en) | 2017-05-11 | 2017-05-11 | A kind of front vehicles detection method in the daytime |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710327872.2A CN107133597A (en) | 2017-05-11 | 2017-05-11 | A kind of front vehicles detection method in the daytime |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107133597A true CN107133597A (en) | 2017-09-05 |
Family
ID=59731481
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710327872.2A Withdrawn CN107133597A (en) | 2017-05-11 | 2017-05-11 | A kind of front vehicles detection method in the daytime |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107133597A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109241916A (en) * | 2018-09-12 | 2019-01-18 | 四川长虹电器股份有限公司 | A kind of system and method for pedestrian's walking safety detection based on smart phone |
| CN110047319A (en) * | 2019-04-15 | 2019-07-23 | 深圳壹账通智能科技有限公司 | Parking position air navigation aid, electronic device and storage medium |
| CN111553388A (en) * | 2020-04-07 | 2020-08-18 | 哈尔滨工程大学 | Junk mail detection method based on online AdaBoost |
| CN112990002A (en) * | 2021-03-12 | 2021-06-18 | 吉林大学 | Traffic signal lamp identification method and system on downhill road and computer readable medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102685516A (en) * | 2011-03-07 | 2012-09-19 | 李慧盈 | Active safety type assistant driving method based on stereoscopic vision |
| CN103269908A (en) * | 2010-12-27 | 2013-08-28 | 丰田自动车株式会社 | Image providing device |
-
2017
- 2017-05-11 CN CN201710327872.2A patent/CN107133597A/en not_active Withdrawn
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103269908A (en) * | 2010-12-27 | 2013-08-28 | 丰田自动车株式会社 | Image providing device |
| CN102685516A (en) * | 2011-03-07 | 2012-09-19 | 李慧盈 | Active safety type assistant driving method based on stereoscopic vision |
Non-Patent Citations (1)
| Title |
|---|
| 金立生 等: ""基于Adaboost算法的日间前方车辆检测"", 《吉林大学学报（工学版）》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109241916A (en) * | 2018-09-12 | 2019-01-18 | 四川长虹电器股份有限公司 | A kind of system and method for pedestrian's walking safety detection based on smart phone |
| CN110047319A (en) * | 2019-04-15 | 2019-07-23 | 深圳壹账通智能科技有限公司 | Parking position air navigation aid, electronic device and storage medium |
| CN111553388A (en) * | 2020-04-07 | 2020-08-18 | 哈尔滨工程大学 | Junk mail detection method based on online AdaBoost |
| CN112990002A (en) * | 2021-03-12 | 2021-06-18 | 吉林大学 | Traffic signal lamp identification method and system on downhill road and computer readable medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8773535B2 (en) | Adaptation for clear path detection using reliable local model updating | |
| CN104573646B (en) | Chinese herbaceous peony pedestrian detection method and system based on laser radar and binocular camera | |
| US8681222B2 (en) | Adaptation for clear path detection with additional classifiers | |
| CN107133597A (en) | A kind of front vehicles detection method in the daytime | |
| US20190392192A1 (en) | Three dimensional (3d) object detection | |
| Wali et al. | Comparative survey on traffic sign detection and recognition: a review | |
| Monteiro et al. | Tracking and classification of dynamic obstacles using laser range finder and vision | |
| CN107463890A (en) | A kind of Foregut fermenters and tracking based on monocular forward sight camera | |
| CN108960074B (en) | Small-size pedestrian target detection method based on deep learning | |
| Romera et al. | A Real-Time Multi-scale Vehicle Detection and Tracking Approach for Smartphones. | |
| CN112381870B (en) | Binocular vision-based ship identification and navigational speed measurement system and method | |
| Fan et al. | Traffic sign detection and classification for Advanced Driver Assistant Systems | |
| US20070223785A1 (en) | Image processor and method | |
| Al-refai et al. | Road object detection using Yolov3 and Kitti dataset | |
| CN108734123A (en) | Highway signs recognition methods, electronic equipment, storage medium and system | |
| Zhang et al. | Moving pedestrian detection based on motion segmentation | |
| Kim et al. | Road surface marking detection for sensor fusion-based positioning system | |
| CN105740783B (en) | Traffic police's detection method and system based on attitude detection Yu clothing feature | |
| Kurnianggoro et al. | Visual perception of traffic sign for autonomous vehicle using k-nearest cluster neighbor classifier | |
| Zakaria et al. | A novel vehicle detection system | |
| Chincholkar et al. | TRAFFIC SIGN BOARD DETECTION AND RECOGNITION FOR AUTONOMOUS VEHICLES AND DRIVER ASSISTANCE SYSTEMS. | |
| Kim et al. | Stereo vision-based improving cascade classifier learning for vehicle detection | |
| Mesmakhosroshahi et al. | Stereo based region of interest generation for pedestrian detection in driver assistance systems | |
| Ajay et al. | Application of Image Processing Techniques for Pothole Detection | |
| Qian et al. | Obstacles extraction using a moving camera |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20170905 |