CN109993134A - A kind of intersection vehicle checking method based on HOG and SVM classifier - Google Patents
A kind of intersection vehicle checking method based on HOG and SVM classifier Download PDFInfo
- Publication number
- CN109993134A CN109993134A CN201910272554.XA CN201910272554A CN109993134A CN 109993134 A CN109993134 A CN 109993134A CN 201910272554 A CN201910272554 A CN 201910272554A CN 109993134 A CN109993134 A CN 109993134A
- Authority
- CN
- China
- Prior art keywords
- image
- gradient
- hog
- svm classifier
- intersection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000001514 detection method Methods 0.000 claims abstract description 31
- 238000000605 extraction Methods 0.000 claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 3
- 239000013598 vector Substances 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000012706 support-vector machine Methods 0.000 description 22
- 238000011835 investigation Methods 0.000 description 5
- 238000007726 management method Methods 0.000 description 5
- 238000007689 inspection Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000012876 topography Methods 0.000 description 2
- 230000017260 vegetative to reproductive phase transition of meristem Effects 0.000 description 2
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of intersection vehicle checking method based on HOG and SVM classifier, the detection recognition method is as follows: the intersection video of acquisition is carried out steady state process, target area after calibrating steady state process, and HOG feature extraction is carried out to the image in figure, detection classification is carried out to image with trained SVM classifier, and constantly rotate image target area, when number of revolutions reaches 5 times, examine whether discrimination meets the requirements, such as it is unsatisfactory for, it is exported the difficult example of error detection as negative sample, re-start SVM training, such as meet the requirements, the target data of acquisition is subjected to rotation inversely processing and obtains former centre coordinate, output center coordinate and image, complete the detection to target.The present invention is able to solve the target detection of the multi-direction vehicle of intersection, and the interference of the influence factors such as vehicle overlapping, intersection channelization facility, building, greening is effectively treated, and improves detection efficiency.
Description
Technical field
The present invention relates to image recognition and object detection fields, more particularly to a kind of be based on HOG (gradient histogram
Figure) and SVM (support vector machines) classifier intersection detection method.
Background technique
With the fast development of economy and society, China's car ownership is significantly risen, and road traffic flow increasingly increases,
Traffic congestion, which has become in work of urban management, one of major issue to be solved.Wherein, most roads generate stifled
The phenomenon that plug the conflict point between vehicle, other than the traffic jam that traffic accident causes on section, intersection is
Vehicle generates the major sources of conflict point, so the basic reason of traffic congestion is intersection, actually crossing traffic function
Caused by traffic obstacle existing for energy, and the investigation of the magnitude of traffic flow is exactly one of traffic planninng and the important foundation of management.Cause
This, the investigation of intersection traffic flow has great importance to the planning and designing and implementation management of road network.
The investigation of the magnitude of traffic flow is one of traffic programme, construction, the key link of management, by the magnitude of traffic flow on road
Investigation, can accurately obtain the magnitude of traffic flow on road network, the essential informations such as vehicle is constituted mention for the prediction and management of traffic condition
For important evidence.At this stage, the investigation of road traffic flow mainly has artificial counting, ground induction coil vehicle detection method, radar inspection
Survey method, video detection method etc..
With the development of machine vision, traditional vehicle checking method is gradually by with convenient for installation and maintenance, low in cost
And the high vehicle checking method based on computer vision of flexibility is replaced.In field of vehicle detection, common method is main
It include: frame differential method, background subtraction, histogram of gradients method (HOG) etc..Frame differential method be by comparing adjacent two frame or
The difference of person's multiple image realizes the detection to target, and scene adaptability is stronger, but more sensitive to ambient noise, dependent on company
The information such as the time interval of continuous frame and car speed；Background subtraction is by carrying out present frame and background model at difference
Reason divides the image into foreground point and background dot using difference result, and the quality of model depends on modeling the Shandong for the background come
Stick, therefore the background area of external condition is changed more sensitive；Method based on direction gradient figure (HOG) be by calculate and
The gradient orientation histogram of statistical picture regional area carrys out constitutive characteristic describer.
The applicable traffic scene of above method is limited, generally can be only applied to the vehicle detection of road single direction, or
Single goal and a small number of targets are detected, the detection to the multi-direction vehicle of intersection can not be suitable for, while
The influence factors such as crossing surrounding environment influence, including intersection channelization facility, building, greening are not accounted for.
Summary of the invention
In order to solve the deficiency that the prior art is not applied for the vehicle detection to the multi-direction vehicle of intersection, this hair
It is bright to provide a kind of intersection vehicle checking method based on HOG and SVM classifier.
To realize the above goal of the invention, the technical solution adopted is that:
A kind of intersection vehicle checking method based on HOG and SVM classifier, comprising the following steps:
Step S1. carries out SVM to existing positive sample and negative sample image zooming-out HOG feature, to the HOG feature after extraction
Training, obtains SVM classifier；
Step S2. obtains traffic scene video to be detected, certain frame image is obtained in traffic scene video to be detected,
And steady state process is carried out to image；
Step S3. goes out the target area in image to the image calibration after steady state process, and divides the image into multiple
Image block carries out HOG feature extraction；
The HOG feature that step S4. is gone out by image zooming-out, detects image using the SVM classifier in step S1
Classification；
Step S5. rotates the target area in image at an angle, it is specified that maximum number of revolutions is M, rotation time
Number is m, and the initial value of m is 1, and every rotation is primary, enables m=m+1, and carry out HOG feature extraction to the target area of image, uses
SVM classifier carries out classification and Detection；
Step S6. judges whether m is greater than M；
(3) if m > M, judge whether discrimination reaches expectation, if not reaching, the difficult example that will identify that is defeated as negative sample
Out, SVM training aids is re-injected into be trained；If discrimination reaches expectation, the target area of acquisition is carried out to rotate inverse place
Reason obtains former centre coordinate, output center coordinate and image, and judges whether there is the abnormal rectangle frame greater than vehicle pixel, if
Have, then excludes abnormal rectangle frame；
(4) if m≤M, step S4 is returned to.
Preferably, the image block of 8 × 8 pixels is divided the image into the step S3.
Preferably, in the step S3 to image block carry out HOG feature extraction the step of include calculate image gradient and
Histogram does normalized to image gradient.
Preferably, specific step is as follows for calculating image gradient and histogram in step S3:
Standardize gamma space: I (x, y)=I (x, y)γ, γ is taken as 0.5, with horizontal edge operator [- 1,0,1] to original image
Convolution algorithm is done, the gradient component G that the direction x i.e. horizontal direction is positive direction is obtainedx；With vertical edge operator [- 1,0,1]TTo original
Image does convolution algorithm, obtains the gradient component G that the direction y i.e. vertical direction is positive directiony；
Using following formula, the gradient at pixel (x, y) is calculated:
Gx(x, y)=H (x+1, y)-H (x-1, y)
Gy(x, y)=H (x, y+1)-H (x, y-1)
G in formulax(x, y), Gy(x, y), H (x, y) respectively indicate horizontal direction gradient at input image pixels point (x, y),
Vertical gradient and pixel value, gradient magnitude and gradient direction at pixel (x, y) are respectively as follows:
Histogram of gradients is constructed to each cell for the image block for being divided into 8 × 8 pixels, area, topography can be obtained
The feature description vectors in domain.
Preferably, do that normalized specific step is as follows in step S3 to image gradient:
Multiple neighbouring Cell cell image blocks are combined into a block block, then find out its gradient direction histogram
Figure vector is normalized, i.e., by histogram vectors using L2-Norm with Hysteresis threshold mode
The maximum value of middle bin value is limited to 0.2 hereinafter, normalizing again again.
Preferably, the value of the maximum number of revolutions M in step S5 is 5.
Compared with prior art, the beneficial effects of the present invention are:
Intersection detection method provided by the invention based on HOG and SVM classifier, is able to solve intersection
The target detection of multi-direction vehicle, and the influence factors such as vehicle overlapping, intersection channelization facility, building, greening are effectively treated
Interference improves detection efficiency.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the present invention.
Fig. 2 is the selected intersection visible image in the embodiment of the present invention.
Fig. 3 is by the target area figure intercepted in the embodiment of the present invention.
Fig. 4 (a), (b), (c), (d), (e), (f) are the image rotated after detection.
Fig. 5 is by multiple rotary treated figure.
Fig. 6 is erroneous detection altimetric image.
Fig. 7 is error detection treated image.
Specific embodiment
The attached figures are only used for illustrative purposes and cannot be understood as limitating the patent；
Below in conjunction with drawings and examples, the present invention is further elaborated.
Embodiment 1
As shown in Figure 1, a kind of intersection vehicle checking method based on HOG and SVM classifier, including following step
It is rapid:
A kind of intersection vehicle checking method based on HOG and SVM classifier, comprising the following steps:
Step S1. carries out SVM to existing positive sample and negative sample image zooming-out HOG feature, to the HOG feature after extraction
Training, obtains SVM classifier；
Step S2. obtains traffic scene video to be detected, certain frame image is obtained in traffic scene video to be detected,
And steady state process is carried out to image；
Step S3. goes out the target area in image to the image calibration after steady state process, and divides the image into multiple
Image block carries out HOG feature extraction；
The HOG feature that step S4. is gone out by image zooming-out, detects image using the SVM classifier in step S1
Classification；
Step S5. rotates the target area in image at an angle, it is specified that maximum number of revolutions is M, rotation time
Number is m, and the initial value of m is 1, and every rotation is primary, enables m=m+1, and carry out HOG feature extraction to the target area of image, uses
SVM classifier carries out classification and Detection；
Step S6. judges whether m is greater than M；
(5) if m > M, judge whether discrimination reaches expectation, if not reaching, the difficult example that will identify that is defeated as negative sample
Out, SVM training aids is re-injected into be trained；If discrimination reaches expectation, the target area of acquisition is carried out to rotate inverse place
Reason obtains former centre coordinate, output center coordinate and image, and judges whether there is the abnormal rectangle frame greater than vehicle pixel, if
Have, then excludes abnormal rectangle frame；
(6) if m≤M, step S4 is returned to.
Preferably, the image block of 8 × 8 pixels is divided the image into the step S3.
Preferably, in the step S3 to image block carry out HOG feature extraction the step of include calculate image gradient and
Histogram does normalized to image gradient.
Preferably, specific step is as follows for calculating image gradient and histogram in step S3:
Standardize gamma space: I (x, y)=I (x, y)γ, γ is taken as 0.5, with horizontal edge operator [- 1,0,1] to original image
Convolution algorithm is done, the gradient component G that the direction x i.e. horizontal direction is positive direction is obtainedx；With vertical edge operator [- 1,0,1]TTo original
Image does convolution algorithm, obtains the gradient component G that the direction y i.e. vertical direction is positive directiony；
Using following formula, the gradient at pixel (x, y) is calculated:
Gx(x, y)=H (x+1, y)-H (x-1, y)
Gy(x, y)=H (x, y+1)-H (x, y-1)
G in formulax(x, y), Gy(x, y), H (x, y) respectively indicate horizontal direction gradient at input image pixels point (x, y),
Vertical gradient and pixel value, gradient magnitude and gradient direction at pixel (x, y) are respectively as follows:
Histogram of gradients is constructed to each cell for the image block for being divided into 8 × 8 pixels, area, topography can be obtained
The feature description vectors in domain.
Preferably, do that normalized specific step is as follows in step S3 to image gradient:
Multiple neighbouring Cell cell image blocks are combined into a block block, then find out its gradient direction histogram
Figure vector is normalized, i.e., by histogram vectors using L2-Norm with Hysteresis threshold mode
The maximum value of middle bin value is limited to 0.2 hereinafter, normalizing again again.
Preferably, the value of the maximum number of revolutions M in step S5 is 5.
Embodiment 2
Step S1: extracting HOG feature to existing vehicle positive sample and negative sample image, special to the HOG after extraction
Sign carries out SVM training, obtains SVM classifier.
Unmanned plane: being hovered over the surface of intersection by step S2, is vertically shot, is intersected to intersection
Mouth video image, and steady state process is carried out, as shown in Figure 1.
Step S3: in the picture by the location information at crossing, object detection area, the inspection of this example middle finger intersection are intercepted out
Survey region.As shown in Figure 2.
Wherein selection includes the image of target to be detected.
Using the image of selection, multiple images block is divided into image and carries out HOG feature extraction, specifically includes following sub-step
It is rapid:
Step S3.1: standardization gamma space: I (x, y)=I (x, y)γ, γ is taken as 0.5.
Step S3.2: convolution algorithm is done to original image with horizontal edge operator [- 1,0,1], obtains the direction x i.e. horizontal direction
For the gradient component G of positive directionx；With vertical edge operator [- 1,0,1]TConvolution algorithm is done to original image, it is i.e. vertical to obtain the direction y
Direction is the gradient component G of positive directiony.Gradient magnitude and gradient direction at pixel (x, y) are respectively as follows:
Step S3.3: it is constructed to the Cell cell for being divided into 8 × 8 pixels in entire image, and for each cell
Gradient orientation histogram obtains the feature description vectors of local image region.
Step S4: on the basis of step 3, carrying out detection classification to image with the SVM classifier after training in step 2,
Fig. 4 (a) is the result of detection for the first time.
Step S5: 30 ° of image rotation will chosen in step 2 carry out step 4 and step 5, repeat 6 times.Fig. 4
(b), Fig. 4 (c), Fig. 4 (d), Fig. 4 (e), Fig. 4 (f) are respectively the result that multiple rotary is examined.(the target area of video image shooting
Domain is cross junction, and the traveling angle of vehicle has diversity, and in order to improve correct verification and measurement ratio, this example chooses rotation angle
It is 30 °).
Step S6: judging whether discrimination reaches to target detection, if not reaching, the difficult example that will identify that is as negative sample
This output re-starts SVM training；If discrimination reaches target detection, the target data of acquisition is subjected to rotation inversely processing
Former centre coordinate, output center coordinate and image are obtained, as shown in Figure 5.
Wherein it should be noted that in step 6 it is possible that error detection to flowering shrubs, as shown in fig. 6, because vehicle
Pixel size does not exceed 20000 squares, and flowering shrubs pixel is much larger than 20000 squares, will be greater than the inspection of 20000 squares
It surveys frame to exclude, as a result as shown in Figure 7.
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair
The restriction of embodiments of the present invention.For those of ordinary skill in the art, may be used also on the basis of the above description
To make other variations or changes in different ways.There is no necessity and possibility to exhaust all the enbodiments.It is all this
Made any modifications, equivalent replacements, and improvements etc., should be included in the claims in the present invention within the spirit and principle of invention
Protection scope within.
Claims (6)
1. a kind of intersection vehicle checking method based on HOG and SVM classifier, which comprises the following steps:
Step S1. carries out SVM instruction to existing positive sample and negative sample image zooming-out HOG feature, to the HOG feature after extraction
Practice, obtains SVM classifier；
Step S2. obtains traffic scene video to be detected, certain frame image is obtained in traffic scene video to be detected, and right
Image carries out steady state process；
Step S3. goes out the target area in image to the image calibration after steady state process, and divides the image into multiple images
Block carries out HOG feature extraction；
The HOG feature that step S4. is gone out by image zooming-out, carries out detection classification to image using the SVM classifier in step S1；
Step S5. rotates the target area in image at an angle, it is specified that maximum number of revolutions is M, and number of revolutions is
The initial value of m, m are 1, and every rotation is primary, enable m=m+1, and carry out HOG feature extraction to the target area of image, use SVM
Classifier carries out classification and Detection；
Step S6. judges whether m is greater than M；
(1) if m > M, judge whether discrimination reaches expectation, if not reaching, the difficult example that will identify that is exported as negative sample,
SVM training aids is re-injected into be trained；If discrimination reaches expectation, the target area of acquisition is subjected to rotation inversely processing and is obtained
To former centre coordinate, output center coordinate and image, and the abnormal rectangle frame greater than vehicle pixel is judged whether there is, if so,
Then abnormal rectangle frame is excluded；
(2) if m≤M, step S4 is returned to.
2. a kind of intersection vehicle checking method based on HOG and SVM classifier according to claim 1, special
Sign is, the image block of 8 × 8 pixels is divided the image into the step S3.
3. a kind of intersection vehicle checking method based on HOG and SVM classifier according to claim 2, special
The step of sign is, carries out HOG feature extraction to image block in the step S3 includes calculating image gradient and histogram, right
Image gradient does normalized.
4. a kind of intersection vehicle checking method based on HOG and SVM classifier according to claim 3, special
Sign is that specific step is as follows for calculating image gradient and histogram in step S3:
Standardize gamma space: I (x, y)=I (x, y)γ, γ is taken as 0.5, rolled up with horizontal edge operator [- 1,0,1] to original image
Product operation obtains the gradient component G that the direction x i.e. horizontal direction is positive directionx；With vertical edge operator [- 1,0,1]TTo original image
Convolution algorithm is done, the gradient component G that the direction y i.e. vertical direction is positive direction is obtainedy；
Using following formula, the gradient at pixel (x, y) is calculated:
Gx(x, y)=H (x+1, y)-H (x-1, y)
Gy(x, y)=H (x, y+1)-H (x, y-1)
G in formulax(x, y), Gy(x, y), H (x, y) respectively indicate horizontal direction gradient at input image pixels point (x, y), vertically
Direction gradient and pixel value, gradient magnitude and gradient direction at pixel (x, y) are respectively as follows:
Histogram of gradients is constructed to each cell for the image block for being divided into 8 × 8 pixels, local image region can be obtained
Feature description vectors.
5. a kind of intersection vehicle checking method based on HOG and SVM classifier according to claim 4, special
Sign is, does that normalized specific step is as follows in step S3 to image gradient:
Multiple neighbouring Cell cell image blocks are combined into a block block, then find out its gradient orientation histogram to
Amount, is normalized using L2-Norm with Hysteresis threshold mode, i.e., will be in histogram vectors
The maximum value of bin value is limited to 0.2 hereinafter, normalizing again again.
6. a kind of intersection vehicle checking method based on HOG and SVM classifier according to claim 5, special
Sign is that the value of the maximum number of revolutions M in step S5 is 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910272554.XA CN109993134A (en) | 2019-04-04 | 2019-04-04 | A kind of intersection vehicle checking method based on HOG and SVM classifier |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910272554.XA CN109993134A (en) | 2019-04-04 | 2019-04-04 | A kind of intersection vehicle checking method based on HOG and SVM classifier |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109993134A true CN109993134A (en) | 2019-07-09 |
Family
ID=67132503
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910272554.XA Pending CN109993134A (en) | 2019-04-04 | 2019-04-04 | A kind of intersection vehicle checking method based on HOG and SVM classifier |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109993134A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111428748A (en) * | 2020-02-20 | 2020-07-17 | 重庆大学 | Infrared image insulator recognition and detection method based on HOG characteristics and SVM |
| CN112598005A (en) * | 2020-12-24 | 2021-04-02 | 深圳市信义科技有限公司 | Character segmentation method for license plate of armed police, storage medium and computer equipment |
| WO2021071422A1 (en) * | 2019-10-10 | 2021-04-15 | Grabtaxi Holdings Pte. Ltd. | Method, data processing apparatus and computer program product for determining road intersections |
| CN113674270A (en) * | 2021-09-06 | 2021-11-19 | 深邦智能科技(青岛)有限公司 | Tire pattern consistency detection system and method |
| CN114519799A (en) * | 2022-02-16 | 2022-05-20 | 复旦大学 | Real-time detection method and system for multi-feature seat state |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011157280A1 (en) * | 2010-06-17 | 2011-12-22 | Valeo Schalter Und Sensoren Gmbh | Method of predetermined road object recognition in images and vehicle driving assistance device |
| CN105160292A (en) * | 2015-07-07 | 2015-12-16 | 深圳市捷顺科技实业股份有限公司 | Vehicle identification recognition method and system |
| CN105787466A (en) * | 2016-03-18 | 2016-07-20 | 中山大学 | Vehicle type fine identification method and system |
| CN106503627A (en) * | 2016-09-30 | 2017-03-15 | 西安翔迅科技有限责任公司 | A kind of vehicle based on video analysis avoids pedestrian detection method |
| CN107808524A (en) * | 2017-11-30 | 2018-03-16 | 中山大学 | A kind of intersection vehicle checking method based on unmanned plane |
-
2019
- 2019-04-04 CN CN201910272554.XA patent/CN109993134A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011157280A1 (en) * | 2010-06-17 | 2011-12-22 | Valeo Schalter Und Sensoren Gmbh | Method of predetermined road object recognition in images and vehicle driving assistance device |
| CN105160292A (en) * | 2015-07-07 | 2015-12-16 | 深圳市捷顺科技实业股份有限公司 | Vehicle identification recognition method and system |
| CN105787466A (en) * | 2016-03-18 | 2016-07-20 | 中山大学 | Vehicle type fine identification method and system |
| CN106503627A (en) * | 2016-09-30 | 2017-03-15 | 西安翔迅科技有限责任公司 | A kind of vehicle based on video analysis avoids pedestrian detection method |
| CN107808524A (en) * | 2017-11-30 | 2018-03-16 | 中山大学 | A kind of intersection vehicle checking method based on unmanned plane |
Non-Patent Citations (2)
| Title |
|---|
| KIKI商: "数字图像处理——行人检测", pages 120 - 121, Retrieved from the Internet <URL:https://blog.csdn.net/qq_18218335/article/details/71791339> * |
| 东方传说: "aaa", Retrieved from the Internet <URL:https://blog.csdn.net/sujunzy666/article/details/41381989> * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021071422A1 (en) * | 2019-10-10 | 2021-04-15 | Grabtaxi Holdings Pte. Ltd. | Method, data processing apparatus and computer program product for determining road intersections |
| CN113841152A (en) * | 2019-10-10 | 2021-12-24 | 格步计程车控股私人有限公司 | Method, data processing device and computer program product for determining a road intersection |
| CN113841152B (en) * | 2019-10-10 | 2022-11-15 | 格步计程车控股私人有限公司 | Method, data processing device and computer program product for determining a road intersection |
| US11663499B2 (en) | 2019-10-10 | 2023-05-30 | Grabtaxi Holdings Pte. Ltd. | Method, data processing apparatus and computer program product for determining road intersections |
| CN111428748A (en) * | 2020-02-20 | 2020-07-17 | 重庆大学 | Infrared image insulator recognition and detection method based on HOG characteristics and SVM |
| CN112598005A (en) * | 2020-12-24 | 2021-04-02 | 深圳市信义科技有限公司 | Character segmentation method for license plate of armed police, storage medium and computer equipment |
| CN113674270A (en) * | 2021-09-06 | 2021-11-19 | 深邦智能科技(青岛)有限公司 | Tire pattern consistency detection system and method |
| CN113674270B (en) * | 2021-09-06 | 2023-11-14 | 深邦智能科技集团(青岛)有限公司 | Tire pattern consistency detection system and method thereof |
| CN114519799A (en) * | 2022-02-16 | 2022-05-20 | 复旦大学 | Real-time detection method and system for multi-feature seat state |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109993134A (en) | A kind of intersection vehicle checking method based on HOG and SVM classifier | |
| CN109118479B (en) | Capsule network-based insulator defect identification and positioning device and method | |
| CN102930296B (en) | A kind of image-recognizing method and device | |
| Luvizon et al. | Vehicle speed estimation by license plate detection and tracking | |
| CN108052624A (en) | Processing Method of Point-clouds, device and computer readable storage medium | |
| CN105260749B (en) | Real-time target detection method based on direction gradient binary pattern and soft cascade SVM | |
| CN104978567B (en) | Vehicle checking method based on scene classification | |
| CN107529659B (en) | Seatbelt wearing detection method, device and electronic equipment | |
| CN106778633B (en) | Pedestrian identification method based on region segmentation | |
| CN104463186B (en) | A kind of target signature detection method and device | |
| CN110197185B (en) | Method and system for monitoring space under bridge based on scale invariant feature transform algorithm | |
| CN108090459A (en) | A kind of road traffic sign detection recognition methods suitable for vehicle-mounted vision system | |
| CN110490839A (en) | The method, apparatus and computer equipment of failure area in a kind of detection highway | |
| CN108596032B (en) | Detection method, device, equipment and medium for fighting behavior in video | |
| CN113724259A (en) | Well lid abnormity detection method and device and application thereof | |
| CN106384359A (en) | Moving target tracking method and television set | |
| US12118640B2 (en) | Hardware accelerator for histogram of oriented gradients computation | |
| CN111079744A (en) | Intelligent vehicle license plate identification method and device suitable for complex illumination environment | |
| CN105631849B (en) | The change detecting method and device of target polygon | |
| CN110188607A (en) | A kind of the traffic video object detection method and device of multithreads computing | |
| CN106446864A (en) | Method for detecting feasible road | |
| Hongquan et al. | Video scene invariant crowd density estimation using geographic information systems | |
| CN105184809A (en) | Moving object detection method and moving object detection device | |
| CN110738229B (en) | Fine-grained image classification method and device and electronic equipment | |
| Laureshyn et al. | Automated video analysis as a tool for analysing road user behaviour |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190709 |