CN109993083A - A kind of vehicle at night knowledge method for distinguishing - Google Patents
A kind of vehicle at night knowledge method for distinguishing Download PDFInfo
- Publication number
- CN109993083A CN109993083A CN201910212457.1A CN201910212457A CN109993083A CN 109993083 A CN109993083 A CN 109993083A CN 201910212457 A CN201910212457 A CN 201910212457A CN 109993083 A CN109993083 A CN 109993083A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- night
- distinguishing
- image
- carried out
- 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 38
- 238000001514 detection method Methods 0.000 claims abstract description 17
- 238000001914 filtration Methods 0.000 claims abstract description 8
- 238000003708 edge detection Methods 0.000 claims abstract description 7
- 230000003628 erosive effect Effects 0.000 claims abstract description 4
- 238000000605 extraction Methods 0.000 claims description 7
- 239000003086 colorant Substances 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- VVNRQZDDMYBBJY-UHFFFAOYSA-M sodium 1-[(1-sulfonaphthalen-2-yl)diazenyl]naphthalen-2-olate Chemical compound [Na+].C1=CC=CC2=C(S([O-])(=O)=O)C(N=NC3=C4C=CC=CC4=CC=C3O)=CC=C21 VVNRQZDDMYBBJY-UHFFFAOYSA-M 0.000 claims description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012407 engineering method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- 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 vehicle at night to know method for distinguishing, comprising the following steps: 1) camera is placed on rearview mirror, obtains video flowing in vehicle traveling process, video flowing is extracted each frame and obtains corresponding picture;2) region division is carried out to above-mentioned picture, establishes several zone lists, extract ROI area-of-interest;3) area-of-interest of acquisition is processed;4) image filtered by YUV, given threshold;5) image obtained is handled by opening operation, first erosion operation, reflation operation;6) gaussian filtering process is carried out again;7) edge detection is finally carried out;8) by the automobile tail light handled above, finally car light is matched, completes vehicle detection at night.The present invention solves the problems, such as that vehicle detection at night method low efficiency and accuracy are low.
Description
Technical field
Automobile identification technology of the present invention field, in particular to a kind of vehicle at night know method for distinguishing.
Background technique
Vehicle detection at night is to judge to whether there is automobile simultaneously in image or video sequence using computer vision technique
It gives and is accurately positioned.The technology can identifies again etc. with pedestrian tracking, pedestrian in conjunction with technologies, be applied to night artificial intelligence system,
The fields such as vehicle DAS (Driver Assistant System), intelligent robot, intelligent video monitoring, human body behavioural analysis, intelligent transportation.
Current vehicle detection at night algorithm is mainly as follows:
1, it is based purely on the vehicle detection at night of HSV
HSV refers to Hue(form and aspect), Saturation(saturation degree) and Value(value), tone H is measured with angle, saturation degree S table
Show degree of the color close to spectrum colour, a kind of color can regard that certain spectrum colour is mixed with white as a result, wherein spectrum as
Ratio shared by color is bigger, and the degree of color close to spectrum colour is just higher, and the saturation degree of color is also just higher;Lightness V indicates face
The bright degree of color, for light source colour, brightness value is related with the bright degree of illuminator;For object color, this value and object
Transmission or reflection ratio is related, and usual value range is that 0%(is black) white to 100%().
2, pedestrian detection classifier is constructed according to a large amount of sample based on the method for statistical learning
The feature of extraction mainly has the information such as gray scale, edge, texture, color, the histogram of gradients of target, and classifier mainly wraps
Include neural network, SVM, Adaboost and the deep learning for being considered as favorite by computer vision now.
3, Faster RCNN and SSD deep learning method carry out vehicle detection.
The shortcomings that above method, is:
1, the problem of background modeling is primarily present at present:
1) car light can be extracted, but street lamp can not filter out;
2) car light adhesion can not be solved the problems, such as;
3) object intensively occurred in image can not identify;
4) effect is bad when having reflective or ponding
2, statistical learning presently, there are difficult point:
1) driving status of vehicle, the background that the type of vehicle is different, complicated and different driving cycles;
2) distribution of the feature extracted in feature space is not compact enough;
3) performance of classifier is affected by training sample;
4) negative sample when off-line training can not cover the case where all true application scenarios;
3, the shortcomings that deep learning method:
1) it is very high to do vehicle accuracy for Faster RCNN and SSD algorithm, but detection speed is very slow, per second to detect 1-2
, it is unsatisfactory for the demand of this scene, and the processor of automotive-type, FPGA can only be used on automobile, operand substantially reduces;
2) current public data collection and the data set of other major colleges and universities are all daytimes, substantially not no night, so adopting
Collection data set has very big difficulty.
Summary of the invention
To solve the problems, such as above-mentioned background technique, the purpose of the present invention is to provide a kind of identifications of vehicle at night
Method solves the problems, such as that vehicle detection at night method low efficiency and accuracy are low in the prior art.
In order to achieve the above objectives, technical scheme is as follows:
A kind of vehicle at night knowledge method for distinguishing, comprising the following steps:
1) camera is placed on rearview mirror, obtains video flowing in vehicle traveling process, video flowing is extracted into each frame and is obtained
Corresponding picture;
2) region division is carried out to above-mentioned picture, establishes several zone lists, extract ROI area-of-interest;
3) area-of-interest of acquisition is processed, first carries out yuv space extraction, Y is expressed as brightness or grayscale value U and V table
It is shown as coloration, YUV describes colors of image and saturation degree, and for the color of specified pixel, brightness is established through RGB input signal
, coloration, that is, shade of color and saturation degree indicate that Cr reflects RGB input signal RED sector and RGB respectively with Cr and Cb
Difference between luminance signals value, what Cb reflected is the same difference of RGB input signal blue portion and rgb signal brightness value,
Cr is individually extracted, it can be seen that filtered out most error;
4) image filtered by YUV, given threshold weed out car light and surface gathered water situation, then by converting with original image
At grayscale image carry out with operation improve accuracy;
5) image obtained is handled by opening operation, and first erosion operation, reflation operation can remove isolated dot, burr
And foot bridge, the location and shape for keeping its total remain unchanged, the ellipse that structural element selection matches with car light ellipse;
6) gaussian filtering process is carried out again;
7) edge detection is finally carried out, edge detection is the apparent point of brightness change in reference numbers image, by handling above,
The stability of vehicle is detected convenient for identifying taillight in flashing;
8) by the automobile tail light handled above, finally car light is matched, completes vehicle detection at night.
Preferably, the video chosen in the step 1) need to embody the operating condition of vehicle at night complexity, video length at least two
Hour.
Preferably, region is carried out according to the ratio that the picture segmentation of 720 resolution ratio is 80 × 80 sizes in the step 2
It divides.
Preferably, the method that brightness is established in the step 3) is that the specific part of rgb signal is superimposed together.
Preferably, the gaussian filtering in the step 6) is a kind of filter of signal, can smoothly be located to signal
Reason, phase, noise are the largest problem to image after treatment, because error can add up the reasons such as transmitting.
Through the above technical solutions, a kind of vehicle at night provided by the invention knows method for distinguishing, propose empty using YUV color
Between and combine some image processing algorithms to exclude some inactive areas (be easy erroneous detection in the area and probability of failure is very big) to come
Guarantee accuracy, meanwhile, and the quantity of vehicle detection will not be reduced.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described.
Fig. 1 is a kind of structural block diagram of vehicle at night knowledge method for distinguishing disclosed in the embodiment of the present invention;
Fig. 2 is a kind of original image and yuv space image pair of vehicle at night knowledge method for distinguishing disclosed in the embodiment of the present invention
Than figure;
Fig. 3 is the given threshold of a kind of vehicle at night knowledge method for distinguishing disclosed in the embodiment of the present invention and the signal with operation
Figure;
Fig. 4 is a kind of vehicle at night knows the opening operation of method for distinguishing disclosed in the embodiment of the present invention and gaussian filtering is matched with lamp
Schematic diagram.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description.
A kind of vehicle at night provided by the invention knows method for distinguishing, as shown in Figure 1, comprising the following steps:
1) camera is placed on rearview mirror, obtains video flowing in vehicle traveling process, video flowing is extracted into each frame and is obtained
Corresponding picture, the video of selection need to embody the operating condition of vehicle at night complexity, video length at least two hour;
2) region division is carried out to above-mentioned picture, establishes several zone lists, extract ROI area-of-interest;According to 720 resolution ratio
Picture segmentation be 80 × 80 sizes ratio carry out region division;
3) area-of-interest of acquisition is processed, first carries out yuv space extraction, Y is expressed as brightness or grayscale value U and V table
It is shown as coloration, YUV describes colors of image and saturation degree, and for the color of specified pixel, brightness is established through RGB input signal
, method is that the specific part of rgb signal is superimposed together, coloration, that is, shade of color and saturation degree, respectively with Cr and Cb
It indicates, Cr reflects difference between RGB input signal RED sector and rgb signal brightness value, and Cb reflection is RGB input
The same difference of signal blue portion and rgb signal brightness value, Cr is individually extracted, it can be seen that filtered out big portion
The error divided;
4) image filtered by YUV, given threshold weed out car light and surface gathered water situation, then by converting with original image
At grayscale image carry out with operation improve accuracy;
5) image obtained is handled by opening operation, and first erosion operation, reflation operation can remove isolated dot, burr
And foot bridge, the location and shape for keeping its total remain unchanged, the ellipse that structural element selection matches with car light ellipse;
6) gaussian filtering process is carried out again, and gaussian filtering is a kind of filter of signal, can be smoothed to signal, is schemed
Phase, noise are the largest problem to picture after treatment, because error can add up the reasons such as transmitting;
7) edge detection is finally carried out, edge detection is the apparent point of brightness change in reference numbers image, by handling above,
The stability of vehicle is detected convenient for identifying taillight in flashing;
8) by the automobile tail light handled above, finally car light is matched, completes vehicle detection at night.
As in Figure 2-4, the result that feature extraction obtains under YUV algorithm, it can be seen that picture passes through a series of figure
The car light extracted after picture processing is more and more clear, and the discrimination of vehicle detection reaches 80% or more.
Under identical algorithms, by engineering method, excludes actual scene and algorithm interference is reached while improving algorithm
The quantity and accuracy of vehicle detection at night.The present invention is based on the vehicle detection at night methods of opencv, for automobile tail light
Post-processing algorithm after extraction algorithm and extraction.
The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present invention.
Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, of the invention
It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one
The widest scope of cause.
Claims (5)
1. a kind of vehicle at night knows method for distinguishing, which comprises the following steps:
1) camera is placed on rearview mirror, obtains video flowing in vehicle traveling process, video flowing is extracted into each frame and is obtained
Corresponding picture;
2) region division is carried out to above-mentioned picture, establishes several zone lists, extract ROI area-of-interest;
3) area-of-interest of acquisition is processed, first carries out yuv space extraction, Y is expressed as brightness or grayscale value U and V table
It is shown as coloration, YUV describes colors of image and saturation degree, and for the color of specified pixel, brightness is established through RGB input signal
, coloration, that is, shade of color and saturation degree indicate that Cr reflects RGB input signal RED sector and RGB respectively with Cr and Cb
Difference between luminance signals value, what Cb reflected is the same difference of RGB input signal blue portion and rgb signal brightness value,
Cr is individually extracted, it can be seen that filtered out most error;
4) image filtered by YUV, given threshold weed out car light and surface gathered water situation, then by converting with original image
At grayscale image carry out with operation improve accuracy;
5) image obtained is handled by opening operation, and first erosion operation, reflation operation can remove isolated dot, burr
And foot bridge, the location and shape for keeping its total remain unchanged, the ellipse that structural element selection matches with car light ellipse;
6) gaussian filtering process is carried out again;
7) edge detection is finally carried out, edge detection is the apparent point of brightness change in reference numbers image, by handling above,
The stability of vehicle is detected convenient for identifying taillight in flashing;
8) by the automobile tail light handled above, finally car light is matched, completes vehicle detection at night.
2. a kind of vehicle at night according to claim 1 knows method for distinguishing, which is characterized in that chosen in the step 1)
Video need to embody the operating condition of vehicle at night complexity, video length at least two hour.
3. a kind of vehicle at night according to claim 1 knows method for distinguishing, which is characterized in that in the step 2) according to
The picture segmentation of 720 resolution ratio is that the ratio of 80 × 80 sizes carries out region division.
4. a kind of vehicle at night according to claim 1 knows method for distinguishing, which is characterized in that established in the step 3) bright
The method of degree is that the specific part of rgb signal is superimposed together.
5. a kind of vehicle at night according to claim 1 knows method for distinguishing, which is characterized in that the Gauss in the step 6)
Filtering is a kind of filter of signal, can be smoothed to signal.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910212457.1A CN109993083A (en) | 2019-03-20 | 2019-03-20 | A kind of vehicle at night knowledge method for distinguishing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910212457.1A CN109993083A (en) | 2019-03-20 | 2019-03-20 | A kind of vehicle at night knowledge method for distinguishing |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109993083A true CN109993083A (en) | 2019-07-09 |
Family
ID=67129707
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910212457.1A Pending CN109993083A (en) | 2019-03-20 | 2019-03-20 | A kind of vehicle at night knowledge method for distinguishing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109993083A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112016552B (en) * | 2020-11-02 | 2021-02-12 | 矿冶科技集团有限公司 | Mixed flotation working condition identification method and system based on foam color |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101969547A (en) * | 2010-07-30 | 2011-02-09 | 新疆宏开电子系统集成有限公司 | Method for processing infrared digital video signal at night |
CN102431530A (en) * | 2011-10-28 | 2012-05-02 | 吉林大学 | Intelligent parking braking and auxiliary starting control method |
CN103208185A (en) * | 2013-03-19 | 2013-07-17 | 东南大学 | Method and system for nighttime vehicle detection on basis of vehicle light identification |
CN104346437A (en) * | 2013-08-09 | 2015-02-11 | 卡尔蔡司显微镜有限责任公司 | Method and data analysis system for semi-automated particle analysis using a charged particle beam |
CN105046222A (en) * | 2015-07-14 | 2015-11-11 | 福州大学 | FPGA-based human face detection and tracking method |
CN107622494A (en) * | 2017-08-28 | 2018-01-23 | 浙江工业大学 | Towards the vehicle detection at night and tracking of traffic video |
US20180262749A1 (en) * | 2015-09-21 | 2018-09-13 | Inuitive Ltd. | Storing Data Retrieved from Different Sensors for Generating a 3-D Image |
CN109447093A (en) * | 2018-10-26 | 2019-03-08 | 上海为森车载传感技术有限公司 | Tail-light detection method based on YUV image |
-
2019
- 2019-03-20 CN CN201910212457.1A patent/CN109993083A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101969547A (en) * | 2010-07-30 | 2011-02-09 | 新疆宏开电子系统集成有限公司 | Method for processing infrared digital video signal at night |
CN102431530A (en) * | 2011-10-28 | 2012-05-02 | 吉林大学 | Intelligent parking braking and auxiliary starting control method |
CN103208185A (en) * | 2013-03-19 | 2013-07-17 | 东南大学 | Method and system for nighttime vehicle detection on basis of vehicle light identification |
CN104346437A (en) * | 2013-08-09 | 2015-02-11 | 卡尔蔡司显微镜有限责任公司 | Method and data analysis system for semi-automated particle analysis using a charged particle beam |
CN105046222A (en) * | 2015-07-14 | 2015-11-11 | 福州大学 | FPGA-based human face detection and tracking method |
US20180262749A1 (en) * | 2015-09-21 | 2018-09-13 | Inuitive Ltd. | Storing Data Retrieved from Different Sensors for Generating a 3-D Image |
CN107622494A (en) * | 2017-08-28 | 2018-01-23 | 浙江工业大学 | Towards the vehicle detection at night and tracking of traffic video |
CN109447093A (en) * | 2018-10-26 | 2019-03-08 | 上海为森车载传感技术有限公司 | Tail-light detection method based on YUV image |
Non-Patent Citations (1)
Title |
---|
宋传学: "油门踏板安全辅助系统开发", 《吉林大学学报(工学版)》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112016552B (en) * | 2020-11-02 | 2021-02-12 | 矿冶科技集团有限公司 | Mixed flotation working condition identification method and system based on foam color |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108765443B (en) | Sign enhancement processing method for self-adaptive color threshold segmentation | |
Sun et al. | HSI color model based lane-marking detection | |
CN109460715A (en) | A kind of traffic lights automatic identification implementation method based on machine learning | |
US11108970B2 (en) | Flicker mitigation via image signal processing | |
CN110688907B (en) | Method and device for identifying object based on night road light source | |
CN106682597A (en) | Method for identifying traffic lights | |
Yang et al. | Real-time traffic sign detection via color probability model and integral channel features | |
US10824885B2 (en) | Method and apparatus for detecting braking behavior of front vehicle of autonomous vehicle | |
CN107315990B (en) | Pedestrian detection algorithm based on XCS-LBP characteristics | |
CN109584258B (en) | Grassland boundary identification method and intelligent mowing device applying same | |
Wonghabut et al. | Traffic light color identification for automatic traffic light violation detection system | |
CN110866932A (en) | Multi-channel tongue edge detection device and method and storage medium | |
CN109948461B (en) | Sign language image segmentation method based on centroid positioning and distance transformation | |
KR100903816B1 (en) | System and human face detection system and method in an image using fuzzy color information and multi-neural network | |
CN112464731A (en) | Traffic sign detection and identification method based on image processing | |
Mandal et al. | Real-time automotive night-vision system for drivers to inhibit headlight glare of the oncoming vehicles and enhance road visibility | |
CN107886033B (en) | Method and device for identifying circular traffic light and vehicle | |
CN111723805B (en) | Method and related device for identifying foreground region of signal lamp | |
CN109993083A (en) | A kind of vehicle at night knowledge method for distinguishing | |
CN112330961B (en) | Traffic indication red, yellow and green light image identification method based on time sequence rule | |
CN113989771A (en) | Traffic signal lamp identification method based on digital image processing | |
Jia et al. | Design of Traffic Sign Detection and Recognition Algorithm Based on Template Matching | |
CN112598674A (en) | Image processing method and device for vehicle and vehicle | |
US10977500B2 (en) | Street marking color recognition | |
Xiang et al. | Research on target detection method based on HSV fusion Gaussian mixture model |
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 |
Application publication date: 20190709 |
|
RJ01 | Rejection of invention patent application after publication |