CN107506760B - Traffic signal detection method and system based on GPS positioning and visual image processing - Google Patents
Traffic signal detection method and system based on GPS positioning and visual image processing Download PDFInfo
- Publication number
- CN107506760B CN107506760B CN201710659729.3A CN201710659729A CN107506760B CN 107506760 B CN107506760 B CN 107506760B CN 201710659729 A CN201710659729 A CN 201710659729A CN 107506760 B CN107506760 B CN 107506760B
- Authority
- CN
- China
- Prior art keywords
- traffic light
- image
- image processing
- traffic
- positioning
- 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
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
-
- 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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
- G06V10/48—Extraction of image or video features by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- 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
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
- G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Atmospheric Sciences (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a traffic signal detection method and a system based on GPS positioning and visual image processing, wherein the method comprises the following steps: s1: triggering a visual image processing process based on the GPS positioning information; s2: acquiring a road front image through a camera device, and respectively sending the road front image into a traffic light frame positioning process and a traffic light signal identification process; s3: detecting whether a traffic light bracket exists in the image, and if so, determining the position of the traffic light bracket; s4: detecting whether a traffic light signal exists in the image, and if so, determining the position and the corresponding signal mode of the traffic light signal; s5: judging whether the position of the lamp body is matched with the position of the lamp holder, and if not, entering the step S6; s6: determining the traffic light signal through pattern recognition, and if the traffic light signal does not exist, returning to the step S2 to acquire the image again; s7: and broadcasting the current traffic signal indication condition by voice. The method has high processing speed and wide covering scene, and can be applied to unmanned driving or blind guiding systems.
Description
Technical Field
The invention relates to the technical field of signal and information processing, in particular to a traffic signal detection method and system based on GPS positioning and visual image processing.
Background
With the development of current unmanned driving and assistant blind guiding systems, people put higher demands on the application of computer vision and image processing technology in traffic. Traffic lights are taken as key traffic signs for indicating the traveling of vehicles and pedestrians, and recently, the traffic lights are widely paid attention by researchers in the information field, and a plurality of traffic light detection and identification algorithms based on computer vision and digital image processing technologies are proposed and are roughly divided into the following four types: methods based on color segmentation, template matching, shape detection, and feature extraction classification. In addition, some intelligent systems based on auxiliary devices, such as GPS, GIS, also provide new approaches to this problem.
The Chinese patent application No. 201410822929.2 discloses a traffic light identification method, which includes the steps of obtaining a binarization result through switching of RGB space gray level binarization and HSV color space binarization methods, extracting a connected domain representing a traffic light signal lamp bright area from the binarization result, designing a classifier, and obtaining a classification identification result of a traffic light target according to the number, color and position characteristics of the extracted connected domains.
The invention patent with application number 201310436336.8 discloses a traffic light automatic identification driving-assisting system, which comprises a driving assisting device arranged on a telegraph pole on which a traffic light is installed, a driving assisting identification instrument, a GPS navigator and a vehicle speed detection unit, wherein the driving assisting device and the driving assisting identification instrument are arranged on a vehicle, and the traffic light is automatically identified by performing two-way communication in a wireless communication mode.
The invention patent with application number 201510181832.2 discloses a traffic light positioning method, which determines whether the current time is day or night by expanding and binarizing a target area where a traffic light is located, and then performs positioning of the traffic light according to different characteristics.
The invention patent with application number 201510208977.7 discloses a traffic light rapid detection algorithm applied to an unmanned automobile, which can improve the detection accuracy of the traffic light by combining color characteristics and sensor data.
The invention patent with application number 201610298509.8 discloses a real-time urban traffic light recognition system based on a monocular vision and GPS combined navigation system, which creates a traffic light map offline through method steps of interactive image marking, camera calibration, three-dimensional position recovery and the like, and then performs color segmentation and shape recognition of traffic lights in an area of interest by utilizing morphological information of the traffic lights. The traffic light identification system is suitable for various different road conditions and scenes, and realizes stable and long-distance detection perception of the traffic light in various environments.
The invention patent with the application number of 201710043802.4 discloses a multi-class traffic light detection method and system based on a prior probability map, which adopts a traffic light detection method based on learning, constructs the prior probability map by utilizing the distribution rule of traffic lights, and carries out self-adaptive setting on a classification threshold value so as to solve the technical problems that the prior art scheme is only applicable to traffic lights with a certain specific shape and has poor anti-jamming capability.
The existing traffic light detection and identification algorithms have problems to a certain extent: (1) the required auxiliary sensing equipment is excessive, and the development platform is not intelligent enough; (2) the surrounding environment can not be effectively understood, such as the judgment and the positioning of the crossroads; (3) the robustness of the related image recognition algorithm in the actual environment is insufficient; (4) there is no relatively complete system framework for smart devices, mostly staying in the laboratory verification phase.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a traffic light detection and identification system based on GPS positioning and computer vision. The system is realized based on the smart phone, and the defects of low accuracy, complex operation, high price and the like of the current traffic light detection and identification system are overcome by deeply researching an image processing algorithm and perfecting the technical framework of the system.
The technical scheme adopted by the invention is as follows: a traffic light detection method based on GPS positioning and visual image processing is characterized by comprising the following steps:
s1: judging whether the current position is in the preset range of the corresponding intersection or not based on the GPS positioning information, if so, entering a step S2 to trigger a visual image processing process;
s2: starting a visual image processing process, acquiring images in front of a road through a camera device, and respectively sending the images into a traffic light frame positioning process and a traffic light signal identification process;
s3: detecting whether a traffic light bracket exists in the image or not through a traffic light frame positioning process, and if so, extracting and determining the position of the bracket in the image through a frame;
s4: detecting whether a traffic light signal exists in the image through a traffic light signal identification process, and if so, determining the position of the lamp body in the image and a corresponding signal mode through color segmentation and Hough transformation;
s5: judging whether the position of the lamp body obtained in the step S4 is matched with the position of the lamp holder obtained in the step S3, if the matching is successful, the traffic light signal detection is accurate, and if the matching is unsuccessful, entering the step S6 to start the posterior process;
s6: adjusting the size of the image according to the position of the lamp holder in the image determined in the step S3, judging whether a traffic light exists in the image through feature extraction, if so, determining a traffic light signal through mode identification, and if not, returning to the step S2 to acquire the image again;
s7: after the specific traffic signal is determined, the current traffic signal indication condition is broadcasted by starting a voice broadcast process.
Further, the positioning process of the traffic light frame in step S3 is specifically performed according to the following steps:
s31: image preprocessing, including image size adjustment and brightness compensation;
s32: segmenting and performing morphological transformation by adopting a self-defined threshold;
s33: positioning the rectangular frame by utilizing the rectangular degree characteristic;
s34: and screening candidate blocks based on the length and width characteristics of the rectangular frame, and finally determining the candidate blocks as the traffic light frame.
Further, the traffic light signal identification process in step S4 is specifically performed according to the following steps:
s41: image preprocessing, including image size adjustment and brightness compensation;
s42: setting a segmentation threshold according to empirical data, performing color segmentation on the RGB image processed in the step S41, and respectively representing the extracted red, green and yellow three-color regions through three binary images generated by the color segmentation;
s43: eliminating interference through Gaussian filtering, median filtering and corrosion expansion operation;
s44: and performing edge detection and Hough circle transformation on the binary image, and finally marking a round traffic light candidate area.
Further, in step S5, whether the lamp body position matches the lamp holder position is determined according to the contact ratio of the lamp body position and the lamp holder position, and the lamp body area is marked as S1The area of the overlapping region of the lamp body and the lamp bracket is S2If the ratio is S1/S2And if the lamp body is positioned in the preset threshold range and the position of the lamp body in the lamp holder corresponds to the correct position, the traffic light signal detection is judged to be accurate.
Further, the posterior process in step S6 includes the following steps:
s61: segmenting the region to be processed according to the position of the lamp holder in the image determined in the step S3;
s62: judging whether the size of the connected region exceeds a threshold value according to the RGB value of the image pixel points, and if the size of the connected region exceeds the threshold value, determining that a traffic light exists;
s63: and extracting characteristic parameters of the connected region, and identifying the traffic light signal mode through a neural network-based classification identifier.
Further, the characteristic parameters input in step S63 include the number of connected regions, the area of each connected region, the average value of RGB components of each connected region, the area ratio of the left and right ends of each connected region, the area ratio of the upper and lower ends of each connected region, the length of the connected region, and the width of the connected region.
Further, the traffic light signal mode identified by the classification identifier comprises three states of red, yellow and green of the disc light; the human body walks in red, yellow and green states; the straight arrow lamp is in red, yellow and green states; the left turn light is in red, yellow and green states, the right turn light is in red, yellow and green states, and the head light is in red, yellow and green states.
Further, after the traffic light signal mode is identified by the classification identifier, the voice broadcasting process broadcasts corresponding voice prompt information.
Based on the statement, the invention also provides a system for realizing the method, which comprises the smart phone, wherein a processor of the smart phone is provided with a visual image processing process control program, positioning information is acquired through a GPS (global positioning system) of the smart phone, a front image is acquired through a camera of the smart phone, and when the visual image processing process in the processor identifies a traffic signal lamp mode, corresponding voice prompt information is output through a voice output module of the smart phone.
By adopting the technical scheme, the method has the following advantages:
the traffic signal detection method and the system based on GPS positioning and visual image processing can accurately detect the position of a traffic light at a crossing and identify a corresponding state; the robustness to interference such as weather is strong; based on GPS intersection positioning, the visual information processing module can realize dynamic operation, thereby reducing resource consumption. According to the invention, a traffic light detection and identification system integrating positioning, image real-time acquisition and processing is built on the smart phone through a computer vision technology and a GPS positioning technology, so that the detection is convenient and quick for a conventional round traffic light, and the intelligent identification is carried out through a classifier for a special traffic light, thereby effectively overcoming the defects of single processing object and narrow application range of the existing system.
Drawings
FIG. 1 is a flow chart of a method of the present invention;
FIG. 2 is a schematic block diagram of the system of the present invention;
FIG. 3 is a partially recognizable traffic light model diagram of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following detailed description is made with reference to the accompanying drawings and specific embodiments, and the description herein does not mean that all the subject matter corresponding to the specific examples set forth in the embodiments is cited in the claims.
As shown in fig. 1-2, a traffic light detection method based on GPS positioning and visual image processing includes the following steps:
s1: judging whether the current position is in the preset range of the corresponding intersection or not based on the GPS positioning information, if so, entering a step S2 to trigger a visual image processing process;
in the step, firstly, a real-time GPS signal is obtained through hardware equipment such as a GPS chip and the like, then positioning software is used for processing the signal to obtain positioning data, finally, the positioning data is matched with a map, and the position of the positioning data is displayed in an electronic map, so that the distance between the positioning data and a corresponding intersection can be calculated. The function of the intersection positioning module is realized by utilizing the Android mobile phone built-in positioning module, the Baidu map API, the Baidu positioning API and the actual road surface data. The Android mobile phone is provided with a GPS module which can be used as hardware equipment for receiving GPS signals; in the aspect of software, the location information is acquired and processed by using a hundred-degree positioning API, and when an application program initiates a positioning request to the positioning API, the positioning API performs positioning according to the actual conditions (such as whether the GPS is turned on, whether a network is connected, whether a signal exists, etc.) of the applied positioning factors (GPS, base station, Wi-Fi signals). In actual use, the GPS is required to be started and is located in an outdoor place where GPS signals can be normally received, a Baidu map API is mainly used for facilitating data display and processing, and topological data of a map is stored in an SQLite database. When the user approaches the intersection, a visual information processing process is triggered.
S2: starting a visual image processing process, acquiring images in front of a road through a camera device, and respectively sending the images into a traffic light frame positioning process and a traffic light signal identification process;
s3: detecting whether a traffic light bracket exists in the image or not through a traffic light frame positioning process, and if so, extracting and determining the position of the bracket in the image through a frame;
in specific implementation, the positioning process of the traffic light frame is specifically performed according to the following steps:
s31: image preprocessing, including image size adjustment and brightness compensation;
s32: segmenting and performing morphological transformation by adopting a self-defined threshold;
s33: positioning the rectangular frame by utilizing the rectangular degree characteristic;
s34: and screening candidate blocks based on the length and width characteristics of the rectangular frame, and finally determining the candidate blocks as the traffic light frame.
The positioning of the traffic light frame is based on the normal or even high exposure image of the actual scene, and the function of extracting the black frame of the traffic light is realized. Because the traffic light frame is black, black is difficult for detecting under the environment that the exposure is low or luminance is low, consequently needs come to carry out the promotion in the aspect of luminance to the image of acquireing according to actual conditions to the detection of frame is convenient for.
The image preprocessing unit receives the intersection image collected by the camera, acquires the basic parameters of the image, adjusts the size of the image and adjusts the exposure of the image (heightens or does not change) according to the actual illumination condition. Usually, the traffic light is set to a certain height, and according to the shooting angle of the camera, the lower half part of the image is usually an invalid image, so that the lower half part of the image needs to be cut off, and only the upper half part of the image needs to be considered. And (3) carrying out RGB histogram equalization on the adjusted image, then carrying out median filtering and image graying processing, and finally outputting and storing a corresponding processing result. The histogram equalization is to perform nonlinear stretching on the image, and redistribute the image pixel values so that the whole gray value of the image is uniformly distributed. RGB histogram equalization is very useful for images where the background and foreground are either too bright or too dark, and can show better detail in either an overexposed or underexposed picture.
And (4) carrying out self-defined threshold segmentation and corrosion expansion operation on the preprocessed image, and then storing a corresponding processing result. The user-defined threshold segmentation algorithm is based on the color characteristics of the black border, and the segmentation threshold is manually set, so that the target and the background are separated. The segmentation method can filter most of interference and is an efficient segmentation method. The threshold value is set to 0.2 to 0.3.
The positioning of the traffic light frame is to position a rectangular frame by utilizing the rectangular degree characteristic, then screen candidate blocks based on the length-width ratio of the rectangular frame, and mark the rest rectangular frame as the traffic light frame.
S4: detecting whether a traffic light signal exists in the image through a traffic light signal identification process, and if so, determining the position of the lamp body in the image and a corresponding signal mode through color segmentation and Hough transformation;
in specific implementation, the traffic light signal identification process is specifically performed according to the following steps:
s41: image preprocessing, including image size adjustment and brightness compensation;
s42: setting a segmentation threshold according to empirical data, performing color segmentation on the RGB image processed in the step S41, and respectively representing the extracted red, green and yellow three-color regions through three binary images generated by the color segmentation;
s43: eliminating interference through Gaussian filtering, median filtering and corrosion expansion operation;
s44: and performing edge detection and Hough circle transformation on the binary image, and finally marking a round traffic light candidate area.
The color segmentation is performed based on the RGB image outputted by the preprocessing, and RGB color segmentation thresholds determined by statistically analyzing the color distribution of a large number of traffic light images at the previous stage are shown in the following table.
Table 1: RGB color segmentation threshold table
Three binary images generated by color segmentation respectively represent the extracted red, green and yellow three-color regions. Eliminating partial interference through Gaussian filtering, median filtering and corrosion expansion operation, and finally outputting and storing the three processed images.
And then, performing edge detection and Hough circle transformation on the binary image, and marking a candidate area of the round traffic light.
The location and status of the traffic light is then determined by combining the circular traffic light area with the previously located traffic light bezel. If the round traffic light is just positioned inside the rectangular frame and the size proportion and the distribution position of the round traffic light are correct, the traffic light is proved to be correctly identified, otherwise, the traffic light is incorrect and needs to be further confirmed. The Hough circle transformation can realize circle detection, and the algorithm comprises the following specific steps:
1. and carrying out edge detection on the input image to obtain boundary points, namely foreground points.
2. If a circle is present in the image, its outline must belong to the foreground point.
3. And (4) carrying out coordinate transformation on the general equation of the circle. And converting the x-y coordinate system into an a-b coordinate system. Written as follows:
(a-x)2+(b-y)2＝r2. A point on the circular boundary in the x-y coordinate system corresponds to a circle in the a-b coordinate system.
4. There are many points on a circular boundary in the x-y coordinate system, and there are many circles corresponding to the a-b coordinate system. Since the points in the original image are all on the same circle, then a, b must also satisfy the equations for all circles in the a-b coordinate system after transformation. The intuitive expression is that the circles corresponding to the plurality of points intersect at one point, and then the intersection point is likely to be the center (a, b).
5. And counting the number of circles at the local intersection points, and taking each local maximum value to obtain the center coordinates (a, b) of the corresponding circle in the original image.
The hough circle transform can reduce the amount of calculation by setting the range of the radius r.
S5: judging whether the position of the lamp body obtained in the step S4 is matched with the position of the lamp holder obtained in the step S3, if the matching is successful, the traffic light signal detection is accurate, and if the matching is unsuccessful, entering the step S6 to start the posterior process;
in specific implementation, in step S5, whether the lamp body is matched with the lamp holder is determined according to the contact ratio between the lamp body and the lamp holder, and the area of the lamp body is recorded as S1The area of the overlapping region of the lamp body and the lamp bracket is S2If the ratio is S1/S2And if the lamp body is within the preset threshold range and the position of the lamp body in the lamp holder corresponds to the correct position (the circular traffic light has relatively fixed position distribution in the frame), the traffic light signal detection is judged to be accurate. The position of the image being easily caused by the angle of captureIn the variation, it cannot be guaranteed that each image is orthographic, so the low threshold is set to 0.1-0.2, and the high threshold is set to 0.5-0.6.
S6: adjusting the size of the image according to the position of the lamp holder in the image determined in the step S3, judging whether a traffic light exists in the image through feature extraction, if so, determining a traffic light signal through mode identification, and if not, returning to the step S2 to acquire the image again;
in specific implementation, the posterior process in step S6 includes the following specific steps:
s61: segmenting the region to be processed according to the position of the lamp holder in the image determined in the step S3;
s62: judging whether the size of the connected region exceeds a threshold value according to the RGB value of the image pixel points, and if the size of the connected region exceeds the threshold value, determining that a traffic light exists;
s63: and extracting characteristic parameters of the connected region, and identifying the traffic light signal mode through a neural network-based classification identifier.
S7: after the specific traffic signal is determined, the current traffic signal indication condition is broadcasted by starting a voice broadcast process.
The characteristic parameters input in step S63 include the number of connected regions, the area of each connected region, the average value of RGB components of each connected region, the area ratio of the left and right ends of each connected region, the area ratio of the upper and lower ends of each connected region, the length of the connected region, and the width of the connected region.
As can be seen in fig. 3, the traffic light signal pattern recognized by the classification recognizer includes three states of red, yellow and green of the disc light; the human body walks in red, yellow and green states; the straight arrow lamp is in red, yellow and green states; the left turn light is in red, yellow and green states, the right turn light is in red, yellow and green states, and the head light is in red, yellow and green states.
After the traffic light signal mode is identified by the classification identifier, the voice broadcasting process broadcasts corresponding voice prompt information.
As can be seen from fig. 2, the embodiment further provides a system for implementing the method, including a smart phone, where a processor of the smart phone is installed with a visual image processing process control program, the processor acquires positioning information through a GPS provided in the smart phone, acquires a front image through a camera provided in the smart phone, and outputs corresponding voice prompt information through a voice output module provided in the smart phone when the visual image processing process in the processor identifies a traffic light mode.
Based on the above design, it can be understood that: the traffic signal detection method and system based on GPS positioning and visual image processing can be realized by adding APP control software on a smart phone directly, have small calculation amount and extremely high processing speed for common round traffic lights, can quickly feed back processing results, can perform classification and identification for unusual abnormal traffic lights by using an intelligent identification algorithm, can effectively detect intersection traffic light signals by combining GPS positioning and voice broadcasting, and provides convenience for unmanned driving and blind navigation.
Finally, while the foregoing is directed to the preferred embodiment of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.
Claims (6)
1. A traffic light detection method based on GPS positioning and visual image processing is characterized by comprising the following steps:
s1: judging whether the current position is in the preset range of the corresponding intersection or not based on the GPS positioning information, if so, entering a step S2 to trigger a visual image processing process;
s2: starting a visual image processing process, acquiring images in front of a road through a camera device, and respectively sending the images into a traffic light frame positioning process and a traffic light signal identification process;
s3: detecting whether a traffic light bracket exists in the image or not through a traffic light frame positioning process, and if so, extracting and determining the position of the bracket in the image through a frame;
s4: detecting whether a traffic light signal exists in the image through a traffic light signal identification process, and if so, determining the position of the lamp body in the image and a corresponding signal mode through color segmentation and Hough transformation;
s5: judging whether the position of the lamp body obtained in the step S4 is matched with the position of the lamp holder obtained in the step S3, if the matching is successful, the traffic light signal detection is accurate, and if the matching is unsuccessful, entering the step S6 to start the posterior process;
s6: adjusting the size of the image according to the position of the lamp holder in the image determined in the step S3, judging whether a traffic light exists in the image through feature extraction, if so, determining a traffic light signal through mode identification, and if not, returning to the step S2 to acquire the image again;
s7: after a specific traffic signal is determined, broadcasting the current traffic signal indication condition by starting a voice broadcasting process;
in step S5, whether the lamp body position matches the lamp holder position is determined according to the contact ratio of the lamp body position and the lamp holder position, and the lamp body area is marked as S1The area of the overlapping region of the lamp body and the lamp bracket is S2If the ratio is S1/S2If the lamp body is within the preset threshold range and the position of the lamp body in the lamp holder is correct, the traffic light signal detection is judged to be accurate;
the positioning process of the traffic light frame in the step S3 is specifically performed according to the following steps:
s31: image preprocessing, including image size adjustment and brightness compensation;
s32: segmenting and performing morphological transformation by adopting a self-defined threshold;
s33: positioning the rectangular frame by utilizing the rectangular degree characteristic;
s34: screening candidate blocks based on the length and width characteristics of the rectangular frame, and finally determining the candidate blocks as a traffic light frame;
the traffic light signal identification process described in step S4 is specifically performed according to the following steps:
s41: image preprocessing, including image size adjustment and brightness compensation;
s42: setting a segmentation threshold according to empirical data, performing color segmentation on the RGB image processed in the step S41, and respectively representing the extracted red, green and yellow three-color regions through three binary images generated by the color segmentation;
s43: eliminating interference through Gaussian filtering, median filtering and corrosion expansion operation;
s44: and performing edge detection and Hough circle transformation on the binary image, and finally marking a round traffic light candidate area.
2. The traffic light detecting method based on GPS positioning and visual image processing as claimed in claim 1, wherein the posterior process in step S6 includes the following steps:
s61: segmenting the region to be processed according to the position of the lamp holder in the image determined in the step S3;
s62: judging whether the size of the connected region exceeds a threshold value according to the RGB value of the image pixel points, and if the size of the connected region exceeds the threshold value, determining that a traffic light exists;
s63: and extracting characteristic parameters of the connected region, and identifying the traffic light signal mode through a neural network-based classification identifier.
3. The traffic light detection method based on GPS positioning and visual image processing according to claim 2, wherein: the characteristic parameters input in step S63 include the number of connected regions, the area of each connected region, the average value of RGB components of each connected region, the area ratio of the left and right ends of each connected region, the area ratio of the upper and lower ends of each connected region, the length of the connected region, and the width of the connected region.
4. The traffic light detection method based on GPS positioning and visual image processing as claimed in claim 3, wherein: the traffic light signal modes identified by the classification identifier comprise three states of red, yellow and green of a disc light; the human body walks in red, yellow and green states; the straight arrow lamp is in red, yellow and green states; the left turn light is in red, yellow and green states, the right turn light is in red, yellow and green states, and the head light is in red, yellow and green states.
5. The traffic light detection method based on GPS positioning and visual image processing as claimed in claim 4, wherein: after the traffic light signal mode is identified by the classification identifier, the voice broadcasting process broadcasts corresponding voice prompt information.
6. A system for implementing the traffic light detection method based on GPS positioning and visual image processing according to claim 1, wherein: the intelligent mobile phone comprises an intelligent mobile phone, wherein a visual image processing process control program is installed in a processor of the intelligent mobile phone, positioning information is obtained through a GPS (global positioning system) of the intelligent mobile phone, a front image is obtained through a camera of the intelligent mobile phone, and when a traffic signal lamp mode is identified by a visual image processing process in the processor, corresponding voice prompt information is output through a voice output module of the intelligent mobile phone.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710659729.3A CN107506760B (en) | 2017-08-04 | 2017-08-04 | Traffic signal detection method and system based on GPS positioning and visual image processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710659729.3A CN107506760B (en) | 2017-08-04 | 2017-08-04 | Traffic signal detection method and system based on GPS positioning and visual image processing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107506760A CN107506760A (en) | 2017-12-22 |
| CN107506760B true CN107506760B (en) | 2020-06-05 |
Family
ID=60690442
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710659729.3A Active CN107506760B (en) | 2017-08-04 | 2017-08-04 | Traffic signal detection method and system based on GPS positioning and visual image processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107506760B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110542931B (en) * | 2018-05-28 | 2021-12-28 | 北京京东乾石科技有限公司 | Traffic light detection method and device, electronic equipment and computer readable medium |
| CN108960200A (en) * | 2018-07-31 | 2018-12-07 | 北京微播视界科技有限公司 | A kind of data processing method and electronic equipment based on intelligent interaction |
| CN111565283A (en) * | 2019-02-14 | 2020-08-21 | 初速度（苏州）科技有限公司 | Traffic light color identification method, correction method and device |
| CN110069986B (en) * | 2019-03-13 | 2021-11-02 | 北京联合大学 | Traffic signal lamp identification method and system based on hybrid model |
| CN110146095B (en) * | 2019-05-23 | 2021-07-20 | 北京百度网讯科技有限公司 | Method and device for navigation of visually impaired people, electronic equipment and computer readable medium |
| CN110415541A (en) * | 2019-07-19 | 2019-11-05 | 重庆长安汽车股份有限公司 | A kind of crossing prevailing state reminding method and system |
| CN110543814B (en) * | 2019-07-22 | 2022-05-10 | 华为技术有限公司 | Traffic light identification method and device |
| CN110598563A (en) * | 2019-08-15 | 2019-12-20 | 北京致行慕远科技有限公司 | Method and device for processing travelling of movable equipment and storage medium |
| CN112766026B (en) * | 2019-11-05 | 2024-02-02 | 大富科技(安徽)股份有限公司 | Traffic light detection method, computer storage medium, vehicle-mounted terminal and vehicle |
| TWI743637B (en) * | 2019-12-27 | 2021-10-21 | 財團法人工業技術研究院 | Traffic light recognition system and method thereof |
| CN112036392B (en) * | 2020-07-22 | 2022-05-03 | 四川长宁天然气开发有限责任公司 | Automatic identification method for states of UPS (uninterrupted Power supply) indicator lamps of production station of shale gas field |
| CN112036391B (en) * | 2020-07-22 | 2023-09-19 | 四川长宁天然气开发有限责任公司 | Electronic inspection method and system based on shale gas field production station |
| CN112587036A (en) * | 2020-12-10 | 2021-04-02 | 苏州阿甘机器人有限公司 | Unmanned sweeper based on machine vision and working method thereof |
| CN113034530A (en) * | 2021-04-02 | 2021-06-25 | 广州绿怡信息科技有限公司 | Mobile phone placement detection method and device |
| CN113177522A (en) * | 2021-05-24 | 2021-07-27 | 的卢技术有限公司 | Traffic light detection and identification method used in automatic driving scene |
| CN113591727A (en) * | 2021-08-03 | 2021-11-02 | 彭刚 | Traffic signal recognition device of distribution robot |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106023623A (en) * | 2016-07-28 | 2016-10-12 | 南京理工大学 | Recognition and early warning method of vehicle-borne traffic signal and symbol based on machine vision |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103778786B (en) * | 2013-12-17 | 2016-04-27 | 东莞中国科学院云计算产业技术创新与育成中心 | A kind of break in traffic rules and regulations detection method based on remarkable vehicle part model |
| CN105389993B (en) * | 2015-12-11 | 2018-02-23 | 余战秋 | The processing and recognition methods of visual traffic signal |
-
2017
- 2017-08-04 CN CN201710659729.3A patent/CN107506760B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106023623A (en) * | 2016-07-28 | 2016-10-12 | 南京理工大学 | Recognition and early warning method of vehicle-borne traffic signal and symbol based on machine vision |
Non-Patent Citations (2)
| Title |
|---|
| 交通信号灯检测与识别算法的研究;黄振威;《中国优秀硕士学位论文全文数据库信息科技辑》;20130215(第02期);第2-25页 * |
| 交通信号灯检测与识别算法研究;葛俊辉;《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》;20150315(第03期);第21-37页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107506760A (en) | 2017-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107506760B (en) | Traffic signal detection method and system based on GPS positioning and visual image processing | |
| CN108804983B (en) | Traffic signal lamp state identification method and device, vehicle-mounted control terminal and motor vehicle | |
| JP5747549B2 (en) | Signal detector and program | |
| JP5423631B2 (en) | Image recognition device | |
| CN108197523B (en) | Night vehicle detection method and system based on image conversion and contour neighborhood difference | |
| CN104616021B (en) | Traffic sign image processing method and device | |
| US20150358549A1 (en) | Image capturing parameter adjustment in preview mode | |
| TWI534764B (en) | Apparatus and method for vehicle positioning | |
| EP2282295A1 (en) | Object recognizing device and object recognizing method | |
| CN101599175B (en) | Detection method for determining alteration of shooting background and image processing device | |
| CN106709412B (en) | Traffic sign detection method and device | |
| CN107886034B (en) | Driving reminding method and device and vehicle | |
| JP2011216051A (en) | Program and device for discriminating traffic light | |
| CN109916415B (en) | Road type determination method, device, equipment and storage medium | |
| CN111881832A (en) | Lane target detection method, device, equipment and computer readable storage medium | |
| CN110084218A (en) | The rainwater distributed data treating method and apparatus of vehicle | |
| JP5200861B2 (en) | Sign judging device and sign judging method | |
| CN110422168B (en) | Lane recognition system and method and automatic driving automobile | |
| CN107886033B (en) | Method and device for identifying circular traffic light and vehicle | |
| CN111046741A (en) | Method and device for identifying lane line | |
| CN107992788B (en) | Method and device for identifying traffic light and vehicle | |
| CN111723805A (en) | Signal lamp foreground area identification method and related device | |
| JP2004086417A (en) | Method and device for detecting pedestrian on zebra crossing | |
| CN112863194A (en) | Image processing method, device, terminal and medium | |
| CN109800693B (en) | Night vehicle detection method based on color channel mixing characteristics |
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 |