CN102768726A - Pedestrian detection method for preventing pedestrian collision - Google Patents
Pedestrian detection method for preventing pedestrian collision Download PDFInfo
- Publication number
- CN102768726A CN102768726A CN2011101165690A CN201110116569A CN102768726A CN 102768726 A CN102768726 A CN 102768726A CN 2011101165690 A CN2011101165690 A CN 2011101165690A CN 201110116569 A CN201110116569 A CN 201110116569A CN 102768726 A CN102768726 A CN 102768726A
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- image
- zone
- detection method
- pedestrians
- 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
Links
Images
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a pedestrian detection method for preventing pedestrian collision. The method comprises the following steps: collecting images in front of a vehicle and pre-processing the images; extracting interesting zones of pedestrians; blocking the images and re-setting the interesting zones; zooming out the to-be-detected zones of each frame of image by one time; pre-localizing the pedestrians in the blocked interesting zones; judging the pre-localized pedestrian zones; and tracking the detected pedestrians. The detection method provided by the invention detects the pedestrians on roads by adopting a pedestrian sorter to fuzz up the individual characteristics of the pedestrians, reduce the effects of the individual differences on the detection result, also reduce the effects of light on the detection result, and enhance the pedestrian detection efficiency; in the meantime, with the adoption of the side pedestrian posture sorter, the pedestrians crossing a road are detected; aiming at the special application scene, the effects of the pedestrians with different postures who do not cross the road on the judgment is reduced; and the effectiveness and the real time property of the algorithm for judging the pedestrians are enhanced.
Description
Technical field
The relevant a kind of pedestrian detection method of the present invention is meant a kind of pedestrian detection method of the prevention pedestrian impact based on machine vision especially.
Background technology
Utilize image processing technique to increase the megatrend that function has become Automobile Design for the automobile supplementary security system.Under the rules of European Union drove, the headstock profile of automobile and structural design had been considered when bumping with the pedestrian, reduce pedestrian's injury and mortality ratio as far as possible.At present, utilize profile and structural design to reduce this method of pedestrian's injures and deaths and obtained certain effect, it is smaller to continue to use the resulting space of improving of this method again.Otherwise, utilize image processing technique to detect the pedestrian and the driver sent prompting even get involved control loop to avoid the pedestrian be the method for relatively keeping forging ahead, and have very big space to improve existing pedestrian protection system.
Existing pedestrian detecting system generally comprises two modules: area-of-interest is cut apart and Target Recognition.The purpose that area-of-interest is cut apart is from image, to extract the window area that possibly comprise the pedestrian to do further checking, to avoid exhaustive search, improves the speed of system.Target Recognition is the core of pedestrian detecting system, and it verifies that to the area-of-interest that obtains to judge wherein whether comprise the pedestrian, its performance has determined accuracy of detection and the robustness that total system can reach.At present, pedestrian detection technology generally has following several kinds of modes: one, and based drive method; Two, based on the method for clear and definite manikin; Three, based on the method for template matches; Four, based on the method for statistical classification.The ultimate principle of above several method and strengths and weaknesses analysis are distinguished as follows:
Based drive method, its principle are to discern the pedestrian through the periodicity of analyzing pedestrian's gait; Advantage is to receive the influence of color, illumination less, relatively robust; Shortcoming is to discern the motion pedestrian, needs multiframe, influences real-time.
Based on the method for clear and definite manikin, its principle is that the clear and definite human parameters model of structure is represented the pedestrian; Advantage is to have clear and definite model, convenient attitude and the occlusion issue handled; Shortcoming is modeling and finds the solution more complicated.
Based on the method for template matches, its principle is through the template representation pedestrian; Advantage is that computing method are simple, and system overhead is little; Shortcoming is to need a lot of templates to tackle the attitude problem, and matching ratio is than time-consuming.
Based on the method for statistical classification, its principle is through sorter the pedestrian to be discerned; Advantage be do not need manual work that quantity of parameters is set, robustness is good; Shortcoming is to need a large amount of training data and cycle of training longer.
Summary of the invention
In view of this, fundamental purpose of the present invention is to provide a kind of automobile that is used for, based on the pedestrian detection method of the prevention pedestrian impact of image processing.
For achieving the above object, the present invention provides a kind of pedestrian detection method that prevents pedestrian impact, and it includes following steps:
(1) gathers vehicle front image and carry out the image pre-service;
(2) extraction pedestrian's area-of-interest;
(3) image is carried out piecemeal, reset area-of-interest;
(4) one times dwindling carried out in the zone to be detected of each two field picture;
(5) in the area-of-interest behind piecemeal the pedestrian is carried out pre-determined bit;
(6) the pedestrian zone of pre-determined bit is judged, with accurate location pedestrian zone;
(7) pedestrian who detects is followed the tracks of.
In the said step (1), through being arranged on the camera collection the place ahead realtime graphic on the automobile, said image pre-service comprises goes sawtooth operation and histogram equalization operation to image.
Said step (2) is based on the positional information that the pedestrian possibly bump in the realtime graphic of collection to be estimated, the image that pedestrian and pedestrian be not in the hazardous location can not occur to part and be not counted in detection, to set pedestrian's area-of-interest.
Said step (3) is that image division is some of equating, the overlapping region is arranged between piece and the piece, each frame input picture is only handled wherein zone, each piece zone of circular treatment.
Utilize first tagsort device pre-determined bit pedestrian zone in pedestrian's area-of-interest in the said step (5); Utilize the second tagsort device that the pedestrian zone of pre-determined bit is screened in the said step (6) and judge; With accurate location pedestrian zone; The tagsort device of the attitude information of pedestrian side under the special scenes that the suitable pedestrian that the said first tagsort device is a training in advance crosses the street, the whole and local tagsort device of the pedestrian that the said second tagsort device is a training in advance.
The said first tagsort device is the rectangular characteristic sorter, and the said second tagsort device is a HOG tagsort device.
Adopt the AdaBoost algorithm in the said step (5), adopt the SVM algorithm in the said step (6).
In said step (5), image being carried out integrogram calculates.
The present invention detects the pedestrian on the road through adopting pedestrian's sorter, has blured the personal feature between the pedestrian, has reduced the influence of individuality difference to testing result, has reduced the influence of illumination to testing result simultaneously, has improved pedestrian's detection efficiency.Through adopting side pedestrian's attitude sorter to detect the pedestrian of crossing road on the road,, reduced the non-influence of crossing road pedestrian differentiation of different attitudes simultaneously, improved algorithm differentiation pedestrian's validity and real-time to the certain applications scene.
Description of drawings
Fig. 1 prevents the process flow diagram of the pedestrian detection method of pedestrian impact for the present invention.
Embodiment
Through embodiment the present invention is done further explain below.
The present invention obtains the realtime graphic of vehicle front through the camera that is installed on the automobile top; Require in image, to extract area-of-interest according to the prevention pedestrian impact; Then a series of Flame Image Process and computing are carried out in the zone of being extracted, realize pedestrian's detection and prevention collision.
Enforcement of the present invention may further comprise the steps:
Step 1; Gather the vehicle front image, gather the place ahead realtime graphic through the camera (for example infrared CCD camera or CMOS camera) that is arranged on the automobile, and image is suitably handled; For example according to the needs of data layout; The image transitions of obtaining is become the single channel gray level image,, image is gone sawtooth operation and histogram equalization operation to meet the requirement of database format.
Step 2 is extracted pedestrian's area-of-interest, estimates based on the positional information that pedestrian in the realtime graphic of gathering possibly bump; The image that pedestrian and pedestrian be not in the hazardous location can not occur to part and be not counted in detection; For example remove the sky and the ground scene of image top and the bottom pixel, and the road both sides scene of image left and right sides partial pixel, thereby setting pedestrian's area-of-interest; Only in more among a small circle, search for the pedestrian; Reduce the Flame Image Process area, thereby reduce data processing amount, improve the algorithm real-time.
Step 3 is carried out piecemeal to image, resets area-of-interest.With image division is some that equate, the overlapping region is arranged between piece and the piece, and the overlapping region is roughly the size of 5-10m pedestrian far away in image.Each frame input picture is only handled wherein zone, each piece of circular treatment.The method has improved the real-time of algorithm greatly.As shown in Figure 1; Pretreated image is divided into two A and B, between piece and the piece overlapping C can be arranged, two zones are respectively piece A+C and piece B+C; Each frame only detect to wherein one carry out; Each piece of cycle detection is followed the tracks of detected pedestrian in the area-of-interest behind the piecemeal, effectively raises the real-time of algorithm.
Step 4 is carried out one times dwindling to the zone to be detected of each two field picture, and zone promptly to be detected dwindles 1/2; Minimum detection frame size is dwindling in proportion also; Under the situation that does not influence testing result, reduced the treatment of picture time, improved the real-time of algorithm; Timing signal reduction detection block size, i.e. expansion is twice.
Step 5 is carried out pre-determined bit to the pedestrian in the area-of-interest behind piecemeal.The attitude information of pedestrian side under the special scenes that crosses the street according to the pedestrian; Training in advance is fit to 12 grade of first tagsort device of this scene pedestrian identification; In the first tagsort device location pedestrian zone of home row man-hour according to training in advance; If no-fix is then returned step 1 to the pedestrian zone.This implementation step specifically comprises:
Adopt the rectangular characteristic template that image is traveled through, according to the first tagsort device that has trained, pre-determined bit pedestrian zone in interesting image regions.For making side pedestrian's the first tagsort device; Use Haar characteristic (being the rectangular characteristic template); Adopt the AdaBoost algorithm to a large amount of side pedestrian's samples and the training of non-pedestrian's background sample; The training on the basis of a large amount of pedestrian's samples and non-pedestrian's background sample data of this first tagsort device is come out, and has stronger universality.
The AdaBoost algorithm can extract in pedestrian's image from a large amount of pedestrian's sample gray level images has distinctive characteristic most, constructs one and has highly accurate pedestrian's first tagsort device; Then, use the pedestrian Haar characteristic that writes down in this first tagsort device that the pedestrian detection area-of-interest is detected and locate, thereby obtain the pedestrian zone of coupling.Utilize the first tagsort device and rectangular characteristic to carry out the location in pedestrian zone, its judgment formula does
Wherein, the expression formula of H (x) expression one-level strong classifier, T representes the Weak Classifier number that the one-level strong classifier is comprised, h
t(x) expression formula of t Weak Classifier of expression, α
tThe weight of representing t Weak Classifier.
By 12 grades cascades, the first tagsort device input picture zone is calculated, if each level is all exported H (x)=1, then representative navigates to the pedestrian zone, if wherein one-level output H (x)=0 is arranged, then representative does not detect the pedestrian zone.
Step 6 is judged the pedestrian zone of pre-determined bit.There is more flase drop in the pedestrian zone of extracting in the step 4; Through these zones being carried out HOG (direction gradient histogram) feature extraction; The second tagsort device according to the pedestrian who has trained is whole and local is judged the characteristic of extracting; Removal flase drop zone, thus accurately locate the pedestrian zone.If do not detect the pedestrian, then return step 1.This implementation step specifically comprises:
1) in order better feature extraction to be carried out in pre-determined bit pedestrian to be checked zone; Can or be contracted to identical size with all pre-determined bit pedestrian to be checked zone expansions, for example the pedestrian's area size after all pre-determined bit is set to the 64*128 fixed size, and the detection block of 64*128 size comprises the piece of 105 16*16 sizes; Piece and interblock have overlapping; Each piece is expressed as the characteristic of 36 dimensions, and whole detection block is described with the proper vector of 3780 dimensions, calculates its 3780 dimensional feature vector.
2) according to the second tagsort device that has trained, the characteristic of each extracted region to be checked is judged, confirmed the pedestrian zone.For making side pedestrian's the second tagsort device; A large amount of side pedestrian's samples and non-pedestrian's background sample are carried out the HOG feature extraction; Adopt the SVM algorithm; Use the libsvm training tool to obtain SVM (Support Vector Machines, SVMs) weights and threshold value, i.e. side pedestrian's the second tagsort device.The second tagsort device provides weight w and threshold value b.Utilize the second tagsort device and the 3780 dimension HOG characteristic x that asked that the zone of pre-determined bit is judged, the detecting pedestrian, its judgment formula does
Wherein, (x representes the proper vector imported for w, the b) weight and the threshold value that provide of the expression second tagsort device, and n is the support vector number of the second tagsort device, (x
i, y
i) i support vector of expression, λ
iBe Lagrangian coefficient.
If f (x)=1, then expression detects and is the pedestrian.
3) the present invention has also taked local feature to judge pedestrian's method, the pedestrian is divided into head, above the waist; Foot areas; Respectively these three zones are combined to specify the training of people from partial row sample, obtain this trizonal sorter respectively, judgement is made in pre-determined bit pedestrian zone.
Step 7 is followed the tracks of the pedestrian who detects.After step 6 is accurately oriented the pedestrian zone, the pedestrian is followed the tracks of the real-time of boosting algorithm greatly, also be simultaneously that of step 3 piecemeal thought replenishes.For example the pedestrian zone of the former frame image that detects is done suitably to enlarge the pedestrian area-of-interest of back as this two field picture, re-use the step 4 method and detect.
When step 5 adopts the rectangular characteristic template that image is traveled through, also can carry out integrogram to image and calculate, it is said that integrogram calculates following formula:
Therefore the traversal of rectangular characteristic template promptly is on resulting integral image, to carry out; To equal in the original image with initial point and this point be the gray scale summation to the rectangular area of angle point to the gray-scale value of every bit on the integral image; Obtain each rectangular characteristic value thus, promptly the eigenwert of rectangular characteristic template is meant rectangle the first half of rectangular characteristic template position and the gray scale difference of the latter half.
The present invention detects the pedestrian on the road through adopting pedestrian's sorter, has blured the personal feature between the pedestrian, has reduced the influence of individuality difference to testing result, has reduced the influence of illumination to testing result simultaneously, has improved pedestrian's detection efficiency.Through adopting side pedestrian's attitude sorter to detect the pedestrian of crossing road on the road,, reduced the non-influence of crossing road pedestrian differentiation of different attitudes simultaneously, improved algorithm differentiation pedestrian's validity and real-time to the certain applications scene.
Above content is to combine concrete embodiment to the further explain that the present invention did, and can not assert that practical implementation of the present invention is confined to these explanations.For the those of ordinary skill of technical field under the present invention, under the prerequisite that does not break away from the present invention's design, can also make some simple deduction or replace, all should be regarded as belonging to protection scope of the present invention.
Claims (8)
1. pedestrian detection method that prevents pedestrian impact is characterized in that it includes following steps:
(1) gathers vehicle front image and carry out the image pre-service;
(2) extraction pedestrian's area-of-interest;
(3) image is carried out piecemeal, reset area-of-interest;
(4) one times dwindling carried out in the zone to be detected of each two field picture;
(5) in the area-of-interest behind piecemeal the pedestrian is carried out pre-determined bit;
(6) the pedestrian zone of pre-determined bit is judged, with accurate location pedestrian zone;
(7) pedestrian who detects is followed the tracks of.
2. the pedestrian detection method of prevention pedestrian impact as claimed in claim 1; It is characterized in that; In the said step (1), through being arranged on the camera collection the place ahead realtime graphic on the automobile, said image pre-service comprises goes sawtooth operation and histogram equalization operation to image.
3. the pedestrian detection method of prevention pedestrian impact as claimed in claim 1; It is characterized in that; Said step (2) is based on the positional information that the pedestrian possibly bump in the realtime graphic of collection and estimates; The image that pedestrian and pedestrian be not in the hazardous location can not occur to part and be not counted in detection, to set pedestrian's area-of-interest.
4. the pedestrian detection method of prevention pedestrian impact as claimed in claim 1; It is characterized in that said step (3) is that image division is some of equating, the overlapping region is arranged between piece and the piece; Each frame input picture is only handled wherein zone, each piece zone of circular treatment.
5. the pedestrian detection method of prevention pedestrian impact as claimed in claim 1; It is characterized in that; Utilize first tagsort device pre-determined bit pedestrian zone in pedestrian's area-of-interest in the said step (5); Utilize the second tagsort device that the pedestrian zone of pre-determined bit is screened in the said step (6) and judge; With accurate location pedestrian zone, the tagsort device of the attitude information of pedestrian side under the special scenes that the suitable pedestrian that the said first tagsort device is a training in advance crosses the street, the whole and local tagsort device of the pedestrian that the said second tagsort device is a training in advance.
6. the pedestrian detection method of prevention pedestrian impact as claimed in claim 5 is characterized in that, the said first tagsort device is the rectangular characteristic sorter, and the said second tagsort device is a HOG tagsort device.
7. the pedestrian detection method of prevention pedestrian impact as claimed in claim 6 is characterized in that, adopts the AdaBoost algorithm in the said step (5), adopts the SVM algorithm in the said step (6).
8. the pedestrian detection method of prevention pedestrian impact as claimed in claim 1 is characterized in that, in said step (5) image is carried out integrogram and calculates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110116569.0A CN102768726B (en) | 2011-05-06 | 2011-05-06 | Pedestrian detection method for preventing pedestrian collision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110116569.0A CN102768726B (en) | 2011-05-06 | 2011-05-06 | Pedestrian detection method for preventing pedestrian collision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102768726A true CN102768726A (en) | 2012-11-07 |
| CN102768726B CN102768726B (en) | 2015-03-11 |
Family
ID=47096120
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110116569.0A Active CN102768726B (en) | 2011-05-06 | 2011-05-06 | Pedestrian detection method for preventing pedestrian collision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102768726B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177263A (en) * | 2013-03-13 | 2013-06-26 | 浙江理工大学 | Image-based automatic detection and counting method for rice field planthopper |
| CN103400113A (en) * | 2013-07-10 | 2013-11-20 | 重庆大学 | Method for detecting pedestrian on expressway or in tunnel based on image processing |
| CN104680124A (en) * | 2013-11-28 | 2015-06-03 | 现代摩比斯株式会社 | Device And Method For Detecting Pedestrains |
| CN105095869A (en) * | 2015-07-24 | 2015-11-25 | 深圳市佳信捷技术股份有限公司 | Pedestrian detection method and apparatus |
| CN105261017A (en) * | 2015-10-14 | 2016-01-20 | 长春工业大学 | Method for extracting regions of interest of pedestrian by using image segmentation method on the basis of road restriction |
| CN106372666A (en) * | 2016-08-31 | 2017-02-01 | 同观科技（深圳）有限公司 | Target identification method and device |
| CN107341441A (en) * | 2017-05-16 | 2017-11-10 | 开易（北京）科技有限公司 | Applied to the pedestrian impact warning system and method in advanced drive assist system |
| CN108764110A (en) * | 2018-05-23 | 2018-11-06 | 大连民族大学 | Recurrence false retrieval method of calibration, system and equipment based on HOG feature pedestrian detectors |
| CN108830210A (en) * | 2018-06-11 | 2018-11-16 | 广东美的制冷设备有限公司 | Human body detecting method and device based on image |
| CN114627651A (en) * | 2022-02-23 | 2022-06-14 | 深圳市锐明技术股份有限公司 | Pedestrian protection early warning method and device, electronic equipment and readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101556695A (en) * | 2009-05-15 | 2009-10-14 | 广东工业大学 | Image matching method |
| CN101887524A (en) * | 2010-07-06 | 2010-11-17 | 湖南创合制造有限公司 | Pedestrian detection method based on video monitoring |
-
2011
- 2011-05-06 CN CN201110116569.0A patent/CN102768726B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101556695A (en) * | 2009-05-15 | 2009-10-14 | 广东工业大学 | Image matching method |
| CN101887524A (en) * | 2010-07-06 | 2010-11-17 | 湖南创合制造有限公司 | Pedestrian detection method based on video monitoring |
Non-Patent Citations (3)
| Title |
|---|
| 叶林 等: "《基于HOG的行人快速检测》", 《计算机工程》 * |
| 吴栓栓: "《基于单目视觉的运动行人检测算法的研究与实现》", 《东北大学硕士学位论文》 * |
| 郭烈: "《基于单目视觉的车辆前方行人检测技术研究》", 《吉林大学博士学位论文》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177263A (en) * | 2013-03-13 | 2013-06-26 | 浙江理工大学 | Image-based automatic detection and counting method for rice field planthopper |
| CN103177263B (en) * | 2013-03-13 | 2016-03-23 | 浙江理工大学 | A kind of rice field plant hopper based on image detects and method of counting automatically |
| CN103400113A (en) * | 2013-07-10 | 2013-11-20 | 重庆大学 | Method for detecting pedestrian on expressway or in tunnel based on image processing |
| CN104680124A (en) * | 2013-11-28 | 2015-06-03 | 现代摩比斯株式会社 | Device And Method For Detecting Pedestrains |
| CN105095869A (en) * | 2015-07-24 | 2015-11-25 | 深圳市佳信捷技术股份有限公司 | Pedestrian detection method and apparatus |
| CN105261017A (en) * | 2015-10-14 | 2016-01-20 | 长春工业大学 | Method for extracting regions of interest of pedestrian by using image segmentation method on the basis of road restriction |
| CN106372666A (en) * | 2016-08-31 | 2017-02-01 | 同观科技（深圳）有限公司 | Target identification method and device |
| CN106372666B (en) * | 2016-08-31 | 2019-07-19 | 同观科技（深圳）有限公司 | A kind of target identification method and device |
| CN107341441A (en) * | 2017-05-16 | 2017-11-10 | 开易（北京）科技有限公司 | Applied to the pedestrian impact warning system and method in advanced drive assist system |
| CN107341441B (en) * | 2017-05-16 | 2020-05-08 | 开易（北京）科技有限公司 | Pedestrian collision warning system and method applied to advanced driving assistance system |
| CN108764110A (en) * | 2018-05-23 | 2018-11-06 | 大连民族大学 | Recurrence false retrieval method of calibration, system and equipment based on HOG feature pedestrian detectors |
| CN108764110B (en) * | 2018-05-23 | 2021-03-23 | 大连民族大学 | Recursive false detection verification method, system and equipment based on HOG characteristic pedestrian detector |
| CN108830210A (en) * | 2018-06-11 | 2018-11-16 | 广东美的制冷设备有限公司 | Human body detecting method and device based on image |
| CN108830210B (en) * | 2018-06-11 | 2021-04-20 | 广东美的制冷设备有限公司 | Human body detection method and device based on image |
| CN114627651A (en) * | 2022-02-23 | 2022-06-14 | 深圳市锐明技术股份有限公司 | Pedestrian protection early warning method and device, electronic equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102768726B (en) | 2015-03-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102768726A (en) | Pedestrian detection method for preventing pedestrian collision | |
| CN102765365B (en) | Pedestrian detection method based on machine vision and pedestrian anti-collision warning system based on machine vision | |
| CN102542289B (en) | Pedestrian volume statistical method based on plurality of Gaussian counting models | |
| Bauer et al. | FPGA-GPU architecture for kernel SVM pedestrian detection | |
| US8750567B2 (en) | Road structure detection and tracking | |
| Satzoda et al. | Multipart vehicle detection using symmetry-derived analysis and active learning | |
| Gerónimo et al. | 2D–3D-based on-board pedestrian detection system | |
| CN108230254B (en) | Automatic detection method for high-speed traffic full lane line capable of self-adapting scene switching | |
| Low et al. | Simple robust road lane detection algorithm | |
| CN103797529B (en) | Three-dimensional body detects device | |
| CN106682586A (en) | Method for real-time lane line detection based on vision under complex lighting conditions | |
| US20070127778A1 (en) | Object detecting system and object detecting method | |
| Kim et al. | Robust detection of non-motorized road users using deep learning on optical and LIDAR data | |
| CN110379168B (en) | Traffic vehicle information acquisition method based on Mask R-CNN | |
| CN102201059A (en) | Pedestrian detection method and device | |
| CN103310194A (en) | Method for detecting head and shoulders of pedestrian in video based on overhead pixel gradient direction | |
| CN105404857A (en) | Infrared-based night intelligent vehicle front pedestrian detection method | |
| CN102682304A (en) | Multi-feature integrated passer-by detection method and device | |
| EP2813973B1 (en) | Method and system for processing video image | |
| Tian et al. | Fast cyclist detection by cascaded detector and geometric constraint | |
| CN104463232A (en) | Density crowd counting method based on HOG characteristic and color histogram characteristic | |
| CN106529404A (en) | Imaging principle-based recognition method for pilotless automobile to recognize road marker line | |
| CN103164958A (en) | Method and system for vehicle monitoring | |
| Xia et al. | Vehicles overtaking detection using RGB-D data | |
| Wang et al. | Pedestrian abnormal event detection based on multi-feature fusion in traffic video |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |