CN109389079B - Traffic signal lamp identification method - Google Patents
Traffic signal lamp identification method Download PDFInfo
- Publication number
- CN109389079B CN109389079B CN201811158223.5A CN201811158223A CN109389079B CN 109389079 B CN109389079 B CN 109389079B CN 201811158223 A CN201811158223 A CN 201811158223A CN 109389079 B CN109389079 B CN 109389079B
- Authority
- CN
- China
- Prior art keywords
- traffic signal
- target
- signal lamp
- traffic
- undetermined
- 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.)
- Active
Links
Images
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- 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
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
- G06V10/422—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation for representing the structure of the pattern or shape of an object therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- 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
- G06V2201/09—Recognition of logos
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a traffic signal lamp identification method, and belongs to the field of image processing. The method comprises the steps of obtaining a traffic scene image; detecting the traffic scene image by using a pre-training model to obtain N pieces of information of undetermined traffic signal lamps; m target traffic signal lamps are selected from the N pending traffic signal lamps; aiming at each target traffic signal lamp, acquiring a maximum connected region of the target traffic signal lamp in a traffic scene image and a binary image corresponding to the maximum connected region to determine the color of the target traffic signal lamp; determining the graph of the target traffic signal lamp according to the binarization graph and the binarization template image corresponding to the maximum communication area; determining the state information of the target traffic signal lamp according to the color and the graph of the target traffic signal lamp; the method solves the problems that the existing method for identifying the state of the traffic signal lamp is complex and is difficult to meet the real-time requirement, and achieves the effects of eliminating interference when identifying a single traffic signal lamp, good real-time performance and simple application.
Description
Technical Field
The embodiment of the invention relates to the field of image processing, in particular to a traffic signal lamp identification method.
Background
With the development of science and technology, the automatic driving technology is becoming more practical. In order to ensure the safety of automatic driving, accurately identifying the traffic signal lamp is an important function that needs to be achieved by an automatic driving system.
In the related art, the traffic signal is mostly recognized by using a conventional image processing method, for example, the state information of the traffic signal is obtained by segmenting and matching the color or the brightness according to the inherent characteristics of the traffic signal. However, the urban environment where the traffic signal lamp is located is complex, and due to the influence of weather, light, obstacles, pedestrians and vehicles, the recognition result is interfered, so that the accuracy of the state information of the recognized traffic signal lamp is not high.
Disclosure of Invention
In order to solve the problems in the prior art, the embodiment of the invention provides a traffic signal lamp identification method. The technical scheme is as follows:
in a first aspect, a traffic signal light identification method is provided, which includes:
acquiring a traffic scene image;
detecting the traffic scene image by using a pre-training model to obtain N pieces of information of undetermined traffic signal lamps; the pre-training model is obtained by training a YOLO V3 model based on a COCO database, and information of one undetermined traffic signal lamp corresponds to one undetermined traffic signal lamp;
according to the information of the undetermined traffic signal lamps and preset position parameters, M target traffic signal lamps are selected from the N undetermined traffic signal lamps; n and M are integers, and N is greater than M;
aiming at each target traffic signal lamp, acquiring a maximum connected region of the target traffic signal lamp in a traffic scene image and a binary image corresponding to the maximum connected region;
taking the color with the largest number of pixel points in the largest connected region as the color of the target traffic signal lamp;
determining a figure of the target traffic signal lamp according to the binarization figure and the binarization template image corresponding to the maximum communication area, wherein the binarization template image comprises a circular lamp, a left-turning arrow, a right-turning arrow and a straight arrow;
determining the state information of the target traffic signal lamp according to the color and the graph of the target traffic signal lamp;
the information of the traffic signal lamp to be determined comprises an abscissa and an ordinate of a top point of the traffic signal lamp to be determined at the upper left of a target area in the traffic scene image, the width and the height of the target area, and a transverse proportion and a longitudinal proportion of the target area in the traffic scene image.
Optionally, according to the information of the pending traffic signal lamps and the preset position parameters, M target traffic signal lamps are selected from the N pending traffic signal lamps, including:
acquiring P undetermined traffic signal lamps from the N undetermined traffic signal lamps, wherein the transverse proportion of the P undetermined traffic signal lamps is in a preset transverse proportion range, and the longitudinal proportion of the P undetermined traffic signal lamps is in a preset longitudinal proportion range; p is an integer, and P is less than N;
aiming at P traffic signal lamps to be determined, sequencing the traffic signal lamps to be determined from large to small in the area of a target area in a traffic scene image;
and taking the front M undetermined traffic signal lamps as target traffic signal lamps.
Optionally, obtaining a maximum connected region of the target traffic signal lamp in the traffic scene image and a binarized image corresponding to the maximum connected region includes:
acquiring a binary image of a target area of a target traffic signal lamp in a traffic scene image;
removing noise points in the binarized image of the target area;
and acquiring the maximum connected region and the binarized image corresponding to the maximum connected region from the binarized image of the target region.
Optionally, the color with the largest number of pixels in the largest connected region is used as the color of the target traffic signal lamp, including:
judging the color of each pixel point in the maximum connected region according to the following formula:
counting the number of pixel points corresponding to each color;
detecting whether the ratio of the maximum number of the pixel points to the total number of the pixel points in the maximum communication area is larger than a preset value or not;
if the ratio of the maximum number of the pixel points to the total number of the pixel points in the maximum communication area is larger than a preset value, taking the color of the pixel points with the maximum number as the color of the target traffic signal lamp;
wherein z isi(x,y)Representing the color of pixel point i with coordinates (x, y).
Optionally, determining the graph of the target traffic signal lamp according to the binarization graph and the binarization template image corresponding to the maximum connected region, including:
aiming at each target traffic signal lamp, converting each binary template image into a template image with the same size as the binary image corresponding to the maximum communication area;
and calculating the square sum of the pixel point number difference between the binarized image corresponding to the maximum connected region and each template image according to the following formula:
taking the square sum of the minimum pixel point number difference and the corresponding binary template image as the graph of the target traffic signal lamp;
wherein D isjIndicating the sum of squares, bw, of the differences in the number of pixel points between the binarized image corresponding to the largest connected region and the ith binarized target imagetarget(x,y)A binarized image, bw, representing the maximum connected region of the target traffic signal lamptemplet(x,y)The ith binarization target image is represented, and (x, y) represents the coordinates of pixel points.
The technical scheme provided by the embodiment of the invention has the following beneficial effects:
the traffic signal lamp in a traffic scene image is primarily identified by adopting a YOLO V3 frame based on a training model of a COCO data set, the result of primary identification is screened according to the regional size and the position information of the traffic signal lamp, the effective region of the traffic signal lamp is protected, the effective region containing the traffic signal lamp is secondarily identified, the color and the graphic characteristics of the traffic signal lamp are obtained, and therefore the state information of the traffic signal lamp is obtained.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a flow diagram illustrating a traffic signal identification method in accordance with an exemplary embodiment;
FIG. 2 is a flow chart illustrating a traffic signal identification method according to another exemplary embodiment;
FIG. 3 is a diagram illustrating a binarized template image according to an exemplary embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
As shown in fig. 1, the method for identifying a traffic signal according to an embodiment of the present invention includes the following steps:
And 102, detecting the traffic scene image by using a pre-training model to obtain N pieces of information of the to-be-determined traffic signal lamp.
The pre-training model is a model trained based on the COCO database using the YOLO V3 model.
And one piece of information of the undetermined traffic signal lamp corresponds to one undetermined traffic signal lamp.
And the confidence coefficient of the information of the undetermined traffic signal lamp is greater than the preset confidence coefficient, and optionally, the preset confidence coefficient is 65%.
N is an integer.
The information of the undetermined traffic signal lamp comprises an abscissa and an ordinate of a top point of the undetermined traffic signal lamp at the upper left of a target area in the traffic scene image, the width and the height of the undetermined traffic signal lamp in the target area in the traffic scene image, and the transverse proportion and the longitudinal proportion of the undetermined traffic signal lamp in the traffic scene image.
The transverse proportion of the traffic signal lamp to be determined in the traffic scene image is the ratio of the abscissa of the top point of the traffic signal lamp to be determined at the upper left of the target area in the traffic scene image to the width of the traffic scene image.
The longitudinal proportion of the traffic signal lamp to be determined in the traffic scene image is the ratio of the ordinate of the top point of the traffic signal lamp to be determined at the upper left of the target area in the traffic scene image to the height of the traffic scene image.
And 103, selecting M target traffic signal lamps from the N undetermined traffic signal lamps according to the undetermined traffic signal lamp information and the preset position parameters.
M is less than N, and M is a positive integer.
And 104, acquiring the maximum connected region of each target traffic signal lamp in the traffic scene image and the binary image corresponding to the maximum connected region.
And 105, taking the color with the largest number of pixel points in the maximum communication area as the color of the target traffic signal lamp.
For example, the colors of the target traffic signal lamp are determined to be red, yellow and green.
And 106, determining the graph of the target traffic signal lamp according to the binarization graph and the binarization template image corresponding to the maximum communication area.
The binary template image comprises a circular lamp, a left-turning arrow, a right-turning arrow and a straight arrow.
And step 107, determining the state information of the target traffic light according to the color and the graph of the target traffic light.
Such as: the color of the target traffic signal lamp is green, the graph is a circular lamp, and the state information of the target traffic signal lamp is a circular green lamp; or the color of the target traffic light is red, the graph is a left-turning arrow, and the state information of the target traffic light is left-turning straight-going.
In summary, in the traffic signal lamp identification method provided by the embodiment of the invention, the traffic signal lamp in the traffic scene image is primarily identified by adopting the YOLO V3 framework based on the training model of the COCO data set, the result of the primary identification is screened according to the size and the position information of the traffic signal lamp area to obtain the effective area for protecting the traffic signal lamp, and the effective area containing the traffic signal lamp is secondarily identified to obtain the color and the graphic characteristics of the traffic signal lamp, so that the state information of the traffic signal lamp is obtained.
Referring to fig. 2, a flow chart of a traffic signal light identification method according to another embodiment of the invention is shown. As shown in fig. 2, the traffic signal light recognition method may include the steps of:
The pre-training model is a model trained based on the COCO database using the YOLO V3 model.
And one piece of information of the undetermined traffic signal lamp corresponds to one undetermined traffic signal lamp. And the confidence coefficient of the information of the undetermined traffic signal lamp is greater than the preset confidence coefficient, and optionally, the preset confidence coefficient is 65%.
N is an integer.
The information of the undetermined traffic signal lamp comprises an abscissa and an ordinate of a top point of the undetermined traffic signal lamp at the upper left of a target area in the traffic scene image, the width and the height of the undetermined traffic signal lamp in the target area in the traffic scene image, and the transverse proportion and the longitudinal proportion of the undetermined traffic signal lamp in the traffic scene image.
The transverse proportion of the traffic signal lamp to be determined in the traffic scene image is the ratio of the abscissa of the top point of the traffic signal lamp to be determined at the upper left of the target area in the traffic scene image to the width of the traffic scene image.
The longitudinal proportion of the traffic signal lamp to be determined in the traffic scene image is the ratio of the ordinate of the top point of the traffic signal lamp to be determined at the upper left of the target area in the traffic scene image to the height of the traffic scene image.
In one example, a traffic scene graph is detected by using a pre-training model, and the obtained N pieces of information of the pending traffic signal lamps are shown as table one.
Watch 1
The left represents the abscissa of the top point of the target area of the traffic signal lamp to be determined in the traffic scene image; top represents the vertical coordinate of the top point of the traffic signal lamp to be determined at the upper left of the target area in the traffic scene image; w represents the width of a target area of the traffic signal lamp to be determined in the traffic scene image; h represents the height of a target area of the traffic signal lamp to be determined in the traffic scene image; lp represents the transverse proportion of a target area of the traffic signal lamp to be determined in the traffic scene image, and lp is left/width of the traffic scene image; tp is top/height of the traffic scene image.
And step 203, acquiring P undetermined traffic signal lamps from the N undetermined traffic signal lamps, wherein the transverse proportion of the P undetermined traffic signal lamps is within a preset transverse proportion range, and the longitudinal proportion of the P undetermined traffic signal lamps is within a preset longitudinal proportion range.
P is a positive integer, and P is less than N.
Since the appearance position of the traffic signal lamp is generally the middle upper part of the traffic scene graph when the vehicle runs normally, a preset transverse proportion range and a preset longitudinal proportion range are set to eliminate targets outside the attention area.
Optionally, the predetermined lateral ratio ranges from 0.2 to 0.8.
Optionally, the predetermined longitudinal proportion ranges from 0 to 0.65.
In one example, the predetermined horizontal ratio range is 0.2-0.8, the predetermined longitudinal ratio range is 0-0.65, and 2 undetermined traffic signal lamps are selected from the 4 undetermined traffic signal lamps shown in the table one, namely the undetermined traffic signal lamp 3 and the undetermined traffic signal lamp 4.
And 204, aiming at the P traffic signal lamps to be determined, sequencing the traffic signal lamps to be determined from large to small in the area of the target area in the traffic scene image.
The area of a target area of the traffic signal lamp to be determined in the traffic scene image is the product of the height of the target area and the width of the target area.
In one example, in the traffic signal information to be specified shown in table one, the area of the target region of the signal lamp 3 is 1560, the area of the target region of the signal lamp 4 is 1701, and the traffic signal lamp to be specified 4 and the traffic signal lamp 3 to be specified are arranged in descending order.
And step 205, taking the front M undetermined traffic signal lamps as target traffic signal lamps.
M is a positive integer. Optionally, M is 4.
When M is larger than P, all the P undetermined traffic signal lamps are used as target traffic signal lamps; and when M is smaller than P, taking the first M traffic signal lamps in the sequenced P undetermined traffic signal lamps as target traffic signal lamps.
In one example, the first 4 pending traffic lights are used as target traffic lights to obtain 2 target traffic lights, and the target traffic light information is shown in table two.
Watch two
The left represents the abscissa of the top point of the target area of the traffic signal lamp to be determined in the traffic scene image; top represents the vertical coordinate of the top point of the traffic signal lamp to be determined at the upper left of the target area in the traffic scene image; w represents the width of a target area of the traffic signal lamp to be determined in the traffic scene image; h represents the height of the target area of the traffic signal lamp to be determined in the traffic scene image.
And step 206, acquiring a binary image of the target traffic signal lamp in the traffic scene image aiming at each target traffic signal lamp.
Optionally, a binarization threshold value is obtained by using a maximum inter-class variance method, and binarization is performed on a target area of the target traffic signal lamp in the traffic scene image.
And step 207, removing noise in the binarized image of the target region.
Optionally, an open operation is applied to remove noise.
And step 208, acquiring the maximum connected region and the binarized image corresponding to the maximum connected region from the binarized image of the target region.
And step 209, judging the color of each pixel point in the maximum communication area according to the formula I.
Wherein z isi(x,y)Representing the color of pixel point i with coordinates (x, y).
When z isi(x,y)When red is obtained, the color of the pixel point i with the coordinate of (x, y) is red; when z isi(x,y)When the pixel point i is yellow, the color of the pixel point i with the coordinate (x, y) is yellow; when z isi(x,y)When green, the color of the pixel point i with coordinates (x, y) is green.
Optionally, the predetermined value is 0.3.
If the ratio of the maximum number of the pixel points to the total number of the pixel points in the maximum connected region is greater than the preset value, indicating that the color is effective, and executing step 212; and if the ratio of the maximum number of the pixel points to the total number of the pixel points in the maximum communication area is not more than a preset value, indicating that the color is invalid, and determining the state of the traffic signal lamp as off.
And step 212, taking the color of the pixel points with the maximum number as the color of the target traffic signal lamp.
It should be noted that, steps 207 to 212 are performed for each target traffic signal lamp.
And step 213, converting each binary template image into a template image with the same size as the binary image corresponding to the maximum connected region for each target traffic signal lamp.
The binarized template image includes a circular lamp, a left-turn arrow, a right-turn arrow and a straight arrow, as shown in fig. 3.
And 214, calculating the square sum of the pixel point quantity difference between the binarized image corresponding to the maximum connected region and each template image according to a formula II.
And calculating the sum of squares of the pixel point number difference between the binarized image corresponding to the maximum connected region of each target traffic signal lamp and each template image.
And step 215, taking the binarized template image corresponding to the sum of squares of the minimum pixel point number difference as the graph of the target traffic signal lamp.
It should be noted that, steps 214 to 215 are performed for each target traffic signal lamp.
And step 216, determining the state information of the target traffic light according to the color and the graph of the target traffic light.
After the color and the graph of each template traffic signal lamp are obtained, sequencing according to the abscissa of the top point of the target traffic signal lamp at the upper left of the target area in the traffic scene image to obtain a target traffic signal lamp combination table; and obtaining an information combination table comprising the number of the corresponding traffic lights according to the target traffic light combination table.
In one example, the target traffic light 1 and the target traffic light 2 in table two are subjected to color and graphic judgment to obtain table three.
Watch III
And obtaining a target traffic signal lamp combination table according to the third table, wherein the table is shown in the fourth table.
Watch four
And obtaining an information combination table comprising the number of the corresponding traffic lights according to the target traffic light combination table, wherein the information combination table is shown in a fifth table.
Watch five
| Left 1 | Left 2 | Left 3 | Left 4 | |
| Single signal lamp | —— | —— | —— | —— |
| Two signal lamps | Left turn green light | Round red light | —— | —— |
| Three signal lamps | —— | —— | —— | —— |
| Four signal lamps | —— | —— | —— | —— |
The traffic signal lamp identification method provided by the embodiment of the invention overcomes the defects of more manual intervention and easy environmental interference of the traditional method due to the manual design characteristics based on deep learning, provides a secondary detection framework based on the traditional model aiming at the defect that the traditional deep learning method needs to specially manufacture a traffic signal lamp training model, and has the advantages of simple implementation and application, better real-time performance and stronger robustness.
It should be noted that: the above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (4)
1. A traffic signal identification method, the method comprising:
acquiring a traffic scene image;
detecting the traffic scene image by using a pre-training model to obtain N pieces of information of the undetermined traffic signal lamps; the pre-training model is obtained by training a YOLO V3 model based on a COCO database, and one piece of information of the undetermined traffic signal lamp corresponds to one undetermined traffic signal lamp;
according to the information of the undetermined traffic signal lamps and preset position parameters, M target traffic signal lamps are selected from the N undetermined traffic signal lamps; n and M are integers, and N is greater than M;
aiming at each target traffic signal lamp, acquiring a maximum connected region of the target traffic signal lamp in the traffic scene image and a binary image corresponding to the maximum connected region;
taking the color with the largest number of pixel points in the maximum communication area as the color of the target traffic signal lamp;
determining the figure of the target traffic signal lamp according to the binarization figure and the binarization template image corresponding to the maximum communication area, wherein the binarization template image comprises a circular lamp, a left-turning arrow, a right-turning arrow and a straight arrow;
determining the state information of the target traffic signal lamp according to the color and the graph of the target traffic signal lamp;
the information of the traffic signal lamp to be determined comprises an abscissa and an ordinate of a top point of the traffic signal lamp to be determined at the upper left of a target area in the traffic scene image, the width and the height of the target area, and a transverse proportion and a longitudinal proportion of the target area in the traffic scene image;
the step of determining the graph of the target traffic signal lamp according to the binarization graph and the binarization template image corresponding to the maximum communication area comprises the following steps:
aiming at each target traffic signal lamp, converting each binaryzation template image into a template image with the same size as the binaryzation image corresponding to the maximum communication area;
and calculating the square sum of the pixel point number difference between the binarized image corresponding to the maximum connected region and each template image according to the following formula:
taking the binary template image corresponding to the sum of squares of the minimum pixel point number difference as the graph of the target traffic signal lamp;
wherein，DjIndicating the sum of squares, bw, of the difference in the number of pixel points between the binarized image corresponding to the largest connected region and the ith binarized target imagetarget(x,y)A binarized image, bw, representing the maximum connected region of the target traffic signal lamptemplet(x,y)The ith binarization target image is represented, and the (x, y) represents the coordinates of the pixel points.
2. The method of claim 1, wherein the extracting M target traffic signals from the N pending traffic signals according to the pending traffic signal information and the preset position parameter comprises:
acquiring P undetermined traffic signal lamps from the N undetermined traffic signal lamps, wherein the transverse proportion of the P undetermined traffic signal lamps is in a preset transverse proportion range, and the longitudinal proportion of the P undetermined traffic signal lamps is in a preset longitudinal proportion range; p is an integer, and P is less than N;
aiming at the P undetermined traffic signal lamps, sequencing the P undetermined traffic signal lamps from large to small according to the area of a target area of the undetermined traffic signal lamps in the traffic scene image;
and taking the front M undetermined traffic signal lamps as target traffic signal lamps.
3. The method according to claim 1, wherein the acquiring a maximum connected region of the target traffic signal lamp in the traffic scene image and a binarized image corresponding to the maximum connected region comprises:
acquiring a binary image of a target area of the target traffic signal lamp in the traffic scene image;
removing noise points in the binarized image of the target area;
and acquiring a maximum connected region and a binarized image corresponding to the maximum connected region from the binarized image of the target region.
4. The method of claim 1, wherein the step of using the color with the largest number of pixels in the largest connected region as the color of the target traffic signal comprises:
judging the color of each pixel point in the maximum communication area according to the following formula:
counting the number of pixel points corresponding to each color;
detecting whether the ratio of the maximum number of the pixel points to the total number of the pixel points in the maximum communication area is larger than a preset value or not;
if the ratio of the maximum number of the pixel points to the total number of the pixel points in the maximum communication area is larger than a preset value, taking the color of the pixel points with the maximum number as the color of the target traffic signal lamp;
wherein z isi(x,y)Representing the color of pixel point i with coordinates (x, y).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811158223.5A CN109389079B (en) | 2018-09-30 | 2018-09-30 | Traffic signal lamp identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811158223.5A CN109389079B (en) | 2018-09-30 | 2018-09-30 | Traffic signal lamp identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109389079A CN109389079A (en) | 2019-02-26 |
| CN109389079B true CN109389079B (en) | 2022-02-15 |
Family
ID=65419105
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811158223.5A Active CN109389079B (en) | 2018-09-30 | 2018-09-30 | Traffic signal lamp identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109389079B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112380973B (en) * | 2020-11-12 | 2023-06-23 | 深兰科技（上海）有限公司 | Traffic signal lamp identification method and system |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102176287A (en) * | 2011-02-28 | 2011-09-07 | 无锡中星微电子有限公司 | Traffic signal lamp identifying system and method |
| CN105913041A (en) * | 2016-04-27 | 2016-08-31 | 浙江工业大学 | Pre-marked signal lights based identification method |
| CN106570494A (en) * | 2016-11-21 | 2017-04-19 | 北京智芯原动科技有限公司 | Traffic signal lamp recognition method and device based on convolution neural network |
| CN106650641A (en) * | 2016-12-05 | 2017-05-10 | 北京文安智能技术股份有限公司 | Traffic light positioning and identification method, device and system |
| CN106781521A (en) * | 2016-12-30 | 2017-05-31 | 东软集团股份有限公司 | The recognition methods of traffic lights and device |
| CN106909937A (en) * | 2017-02-09 | 2017-06-30 | 北京汽车集团有限公司 | Traffic lights recognition methods, control method for vehicle, device and vehicle |
| CN107038420A (en) * | 2017-04-14 | 2017-08-11 | 北京航空航天大学 | A kind of traffic lights recognizer based on convolutional network |
| CN107527511A (en) * | 2016-06-22 | 2017-12-29 | 杭州海康威视数字技术股份有限公司 | A kind of intelligent vehicle driving based reminding method and device |
| CN107704853A (en) * | 2017-11-24 | 2018-02-16 | 重庆邮电大学 | A kind of recognition methods of the traffic lights based on multi-categorizer |
| CN108108761A (en) * | 2017-12-21 | 2018-06-01 | 西北工业大学 | A kind of rapid transit signal lamp detection method based on depth characteristic study |
-
2018
- 2018-09-30 CN CN201811158223.5A patent/CN109389079B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102176287A (en) * | 2011-02-28 | 2011-09-07 | 无锡中星微电子有限公司 | Traffic signal lamp identifying system and method |
| CN105913041A (en) * | 2016-04-27 | 2016-08-31 | 浙江工业大学 | Pre-marked signal lights based identification method |
| CN107527511A (en) * | 2016-06-22 | 2017-12-29 | 杭州海康威视数字技术股份有限公司 | A kind of intelligent vehicle driving based reminding method and device |
| CN106570494A (en) * | 2016-11-21 | 2017-04-19 | 北京智芯原动科技有限公司 | Traffic signal lamp recognition method and device based on convolution neural network |
| CN106650641A (en) * | 2016-12-05 | 2017-05-10 | 北京文安智能技术股份有限公司 | Traffic light positioning and identification method, device and system |
| CN106781521A (en) * | 2016-12-30 | 2017-05-31 | 东软集团股份有限公司 | The recognition methods of traffic lights and device |
| CN106909937A (en) * | 2017-02-09 | 2017-06-30 | 北京汽车集团有限公司 | Traffic lights recognition methods, control method for vehicle, device and vehicle |
| CN107038420A (en) * | 2017-04-14 | 2017-08-11 | 北京航空航天大学 | A kind of traffic lights recognizer based on convolutional network |
| CN107704853A (en) * | 2017-11-24 | 2018-02-16 | 重庆邮电大学 | A kind of recognition methods of the traffic lights based on multi-categorizer |
| CN108108761A (en) * | 2017-12-21 | 2018-06-01 | 西北工业大学 | A kind of rapid transit signal lamp detection method based on depth characteristic study |
Non-Patent Citations (5)
| Title |
|---|
| 基于图像处理的交通信号灯识别方法;武莹等;《交通信息与安全》;20111231;第29卷(第3期);第51-54页 * |
| 基于图像的交通信号灯识别;王涛;《万方数据知识服务平台》;20160831;第3-4章 * |
| 基于深度学习的交通信号灯检测及分类方法;王莹等;《汽车实用技术》;20180915(第17期);第89-91页 * |
| 智能车交通灯识别;李广亮等;《杭州电子科技大学学报》;20140531;第34卷(第3期);第123-126页 * |
| 缩微交通环境下交通灯识别方法研究;刘忆萱;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20160815(第8期);C034-313 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109389079A (en) | 2019-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103077384B (en) | A kind of method and system of vehicle-logo location identification | |
| CN101408942B (en) | Method for locating license plate under a complicated background | |
| CN102693423B (en) | One pinpoint method of car plate under intense light conditions | |
| CN103383733B (en) | A kind of track based on half machine learning video detecting method | |
| CN105488454A (en) | Monocular vision based front vehicle detection and ranging method | |
| CN109726717B (en) | Vehicle comprehensive information detection system | |
| CN111027511B (en) | Remote sensing image ship detection method based on region of interest block extraction | |
| CN107492094A (en) | A kind of unmanned plane visible detection method of high voltage line insulator | |
| CN107016362B (en) | Vehicle weight recognition method and system based on vehicle front windshield pasted mark | |
| CN109215364B (en) | Traffic signal recognition method, system, device and storage medium | |
| CN104050447A (en) | Traffic light identification method and device | |
| CN108256467B (en) | Traffic sign detection method based on visual attention mechanism and geometric features | |
| CN104182754A (en) | Rural resident point information extraction method based on high-resolution remote-sensing image | |
| CN109087363B (en) | HSV color space-based sewage discharge detection method | |
| CN103729863B (en) | Traffic lights automatic location based on autonomous learning is known method for distinguishing | |
| CN110688907A (en) | Method and device for identifying object based on road light source at night | |
| CN105809149A (en) | Lane line detection method based on straight lines with maximum length | |
| CN107563301A (en) | Red signal detection method based on image processing techniques | |
| US10726277B2 (en) | Lane line detection method | |
| CN105117726A (en) | License plate positioning method based on multi-feature area accumulation | |
| CN103093200A (en) | Algorithm for quickly and accurately locating plate number of image | |
| CN108648210B (en) | Rapid multi-target detection method and device under static complex scene | |
| CN109389079B (en) | Traffic signal lamp identification method | |
| CN108416284A (en) | A kind of dividing method of traffic lights | |
| CN113989771A (en) | Traffic signal lamp identification method based on digital image processing |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |