WO2012081963A1

WO2012081963A1 - System and method for traffic violation detection

Info

Publication number: WO2012081963A1
Application number: PCT/MY2011/000145
Authority: WO
Inventors: Sheau Wei Chau; Yen San Yong; Ching Hau Chan; Hafrizan Bin Zakaria Khairil; Siu Jing Then; Hock Woon Hon
Original assignee: Mimos Bhd
Current assignee: Mimos Bhd
Priority date: 2010-12-15
Filing date: 2011-06-24
Publication date: 2012-06-21
Anticipated expiration: 2013-06-15
Also published as: MY180285A

Abstract

The present invention relates to a system and method for traffic violation detection, more particularly to a method to detect traffic violation from differentiating between normal traffic behaviors and illegal traffic behaviors with its associated system. The system comprises a detector (101), a capturing means (102), a processor (103), and a storage means (104), for providing training to the system of traffic violation detection, and operating the trained system for traffic violation detection of illegal driving behavior profiles, such illegal U-turns, illegal right turn, illegal left turn and so on.

Description

SYSTEM AND METHOD FOR TRAFFIC VIOLATION DETECTION

TECHNICAL FIELD

The technical field of the invention relates to a system and method for traffic violation detection, more particularly to a system and method to detect traffic violation from differentiating between normal traffic behaviors and illegal traffic behaviors.

BACKGROUND OF INVENTION

The ever-growing number of vehicles on the road has indeed caused an irritable traffic flow where the control of traffic seem unmanageable with drivers tend to violate the traffic rules in order to get to their destination in the shortest time possible. These unscrupulous drivers usually violate traffic rules at their own convenience and risk while leaving other drivers in despicable danger, such as making an illegal U-turn, illegal right turn, illegal left turn, or even failing to bring their vehicle to a halt at compulsory stops. More significantly, the increasing number of vehicles on the road has tremendously increased traffic violation activities with a consequent of mounting the number of road accidents. Moreover, with the limitation of enforcement personnel manpower to monitor areas of traffic violation, the problem persists longer with road accidents and number of traffic offenders on the rise. The problem becomes even more complicated at crossroads where there is a fine line between making U-turn and making right turn.

Several prior arts have disclosed system and method for detecting traffic violation. One of the prior art, WO/2003/088177 disclosed a system and a method for recording a traffic violation in which a vehicle is involved, comprises detecting the incident using a radar trap that emits a beam of radar waves at a predetermined angle relative to the longitudinal axis of the a road as to detect speed of a vehicle that drives through the radar beam by the vehicle reflecting some of the radar waves back, making at least one record of the detected incident, and searching for and reading from the record a license plate of the vehicle involved in the incident. US 4847772 (A) disclosed a traffic detection and monitoring equipment, in which infrared or visible images of highway scenes are processed by digital computing means to determine vehicle presence, passage, measure various traffic parameters and facilitate traffic surveillance and control. US2002072847 (Al) disclosed method and apparatus for monitoring traffic events using vision-based recognition techniques, wherein the methods suggest establishing at least one rule defining said vehicular traffic event, processing at least one image of vehicular traffic to identify a condition, and performing an action item if said rule is satisfied for traffic violations selected from the group consisting of an illegal turn, an excessive speed, and failure to stop at a stop sign. US 2004054513 (Al) disclosed a system for detecting and filtering nonviolation events in order to more effectively allocate resources within traffic violation detection and recording system. Methods that are adapted in this patent includes a virtual violation line corresponding to a location at an intersection that if crossed by a vehicle entering said intersection during a red light indication of a traffic signal, is determined to be a violation by the vehicle.

These prior arts mostly concentrated on the violation of vehicle speed limits and traffic red-light violation. In view of the present invention, there is a need for system and method for detecting traffic violation at areas where prohibition of illegal traffic conducts require accurate detection and appropriate enforcement to ensure that traffic violators are brought to the face of justice, resulting in a safer driving environment for guiltless road users.

SUMMARY OF INVENTION

A system and a method for traffic violation detection in accordance with the present invention are disclosed herein. The object of the present invention is to provide a system comprising a detector, a capturing means, a processor, and a storage means for effectively tracking the presence of a vehicle and detecting any forms of traffic violation of the vehicle, wherein the detector transmits a detection signal to the processor for processing and involving a mutual process with the storage means and the capturing means. Preferably, the system for traffic violation detection comprises a detector that includes a plurality of sensor arranged in a single line, arrayed for tracking the presence of at least a vehicle in which the sensors are photosensitive elements, wherein said processor receives digital input from the detector for motion information detection.

Preferably, the system for traffic violation detection comprises a processor for processing information received from the detector, and mutually processing information with the capturing means and the storage means, for providing continuous information processing between the detector, capturing means, and the storage means.

Preferably, the system for traffic violation detection comprises a capturing means for capturing image information of at least a vehicle, and storage means for retrieving and storing information of said detector, said processor, and said capturing means. Preferably, the method for traffic violation detection, wherein training the system for traffic violation detection comprises creating various driving behavior profiles of predetermined legal and illegal driving behavior profiles from an analysis performed within various drive trails as determined by array of single lines of a plurality of sensors, hence assigning each drive trail to a corresponding driving behavior profile type.

Preferably, the method for traffic violation detection, wherein operating the trained system for traffic violation detection comprises continuously extracting time and position information from an array of single lines of a plurality of sensors, processing the time and position information for identifying a location and temporal motion pattern of the detected vehicle, comparing the location and temporal motion pattern information with the driving behavior profile created for identifying driving behavior type of the detected vehicle, thus triggering alarm and capturing an image information of the vehicle for driving profile of said vehicle matching an illegal driving behavior type.

Therefore, the present invention addresses the inevitable occurrences of traffic rule violations on a street, highway, crossroads or even road intersections with the presence of traffic lights, where manpower of the enforcement authorities are off limits and costs of workforce are out of bounds. The present invention is capable of alleviating traffic crimes and road accidents, wherein the system and method of the present invention overlooks driving behaviors of road users, detects traffic offenders, captures appropriate information, and reports to the authorities. Typically, the dependency of the method for traffic violation detection in the present invention is the processing methods of detecting an illegal driving behavior.

BRIEF DESCRIPTION OF DRAWINGS

Figure 1: Illustrates a system for traffic violation detection in accordance with the present invention.

Figure 2: Illustrates a detector for traffic violation detection corresponding to a graph representation in accordance with the present invention.

Figure 3: Illustrates a process flow of a method for training a system for traffic violation detection in accordance with the present invention.

Figure 4: Illustrates a process flow of a method for operating a trained system for traffic violation detection in accordance with the present invention. Figure 5 (a): Illustrates a driving behavior corresponding to a type of driving behavior profile in accordance with the present invention. Figure 5 (b): Illustrates a driving behavior corresponding to a type of driving behavior profile in accordance with the present invention. Figure 6 (a): Illustrates a driving behavior corresponding to a type of driving behavior profile in accordance with the present invention. Figure 6 (b): Illustrates a driving behavior corresponding to a type of driving behavior profile in accordance with the present invention.

DETAILED DESCRIPTION OF EMBODIMENTS

Described below are preferred embodiments of the present invention with reference to the accompanying drawings. Each of the following preferred embodiments describes an example not limiting in any aspect. Referring to Figure 1 , one embodiment of a system for traffic violation detection in accordance with the present invention, wherein the system comprises a detector (101) for detecting presence of at least a vehicle, a capturing means (102) for capturing image information of at least a vehicle, a processor (103) for processing information received from the detector (101) and mutually operating with the capturing means

(102) , and a storage means (104) mutually operating with the processor (103) for retrieving and storing information of said detector (101), said processor (103) and said capturing means (102). Source of initialization of the entire system solely depends on information received from the detector (101) as shown in the figure, whereby information acquired from the detector (101) is transmitted directly to a processor (103) through any means of mediation for the processor (103) to read the transmitted information. The processor

(103) preferably receives digital input from the detector (101) for motion information detection, as well involves processing information received from the detector (101), and mutually processing information with the capturing means (102) and the storage means (104), for providing continuous information processing between the detector (101), capturing means (102), and the storage means (104). The capturing means (102) is preferably a traffic camera to well capture a snapshot of a violating vehicle upon receiving information from the processor (103) of an instruction to take a snapshot towards producing image information for analysis. The storage means (104) is preferably a storage device to store information typically processed from the processor (103). Information that is stored in the storage means (104) consists of processed information that is received from the detector (101) and the capturing means (102). The storage means (104) also plays an important role for retrieving stored information in the event of detecting a traffic violation.

Referring to Figure 2, there is shown a detector (101) device for traffic violation detection corresponding to a graph representation (201) in accordance with the present invention, the detector (101) includes a plurality of sensors (105) arranged in a single line, arrayed for tracking presence of at least a vehicle. The plurality of sensors (105) are preferably photosensitive elements to detect the presence of a vehicle driven through the detector (101), in which these photosensitive elements tracks the drive trail of at least a vehicle as the vehicle drives through each and every single lines of a plurality of sensors (105). Usually the detector (101) device is operated using a detector circuit, whereby analogue signals detected by the photosensitive elements are converted to voltage readings which is then used as digital input for motion information detection which will then used as a basis for decision. Alternatively, the detector (101) can also be constructed using a line-scan sensor and optical fibers, wherein the photosensitive elements on a line-scan sensor of either charge-coupled device (CCD) or complementary metal-oxide-semiconductor (CMOS) are noticeably extended when coupled with the optical fibers. This coupling method has the opening end of the optical fiber used as an extended light-sensing device, and the line-scan sensor can directly read the light level as if it is directly sensing from the sensor. This provides flexibility for light sensing, as optical fiber is bendable while preserving its light conveying capability.

In the same context, the tracked drive trail of a vehicle is only logically true for vehicle that crosses every single lines of a plurality of sensors (105) constructed within the detector (101), whereby every adjacent single lines of a plurality of sensors (105) has to transmit a signal of presence of the vehicle sequentially, failing which would result in inconsistency of sequence of a drive trail of said vehicle and would be treated as a detection of a new vehicle.

The corresponding graph representation (201) of the detector (101) merely represents a drive trail of a detected vehicle when the detector (101) is supposedly in operation, in which the y-axis of the graph representation (201) indicates location variation of the detected vehicle, and the x-axis of the graph representation (201) indicates time taken for the detected vehicle to cross each and every single lines of a plurality of sensors (105). Similarly, the y-axis of the graph representation (201) represents the length of sensors (105) arranged in a single line, indicating the location variation. Any signal detected along the line of sensors will indicate a position of a point generated along the y-axis of the graph representation (201). The x-axis of the graph representation (201) represents the time sequence of a vehicle detected for each and every single lines of a plurality of sensors (105), where a successful detection of a single line of a plurality of sensors (105) will indicate a position of a point generated along the x-axis of the graph representation (201). In particular, a method for traffic violation detection as adapted in the present invention involves two main sub divisional methods, that is one for training a system for traffic violation detection and another for operating the trained system for traffic violation detection. The entire process flow for traffic violation detection will be further elaborated in the following descriptions herein.

Referring now to Figure 3, there is illustrated a process flow of a method for training a system for traffic violation detection in accordance with the present invention (202), wherein the method for training the system is initialized with receiving a series of timing and location information from a plurality of sensors (105) arranged in single line, arrayed for creating instances of references to be referred during the training, and determining an initial detection of a vehicle for indicating the initial time and position of the detected vehicle by a first line sensors (107), thus recording the initial time and position of the detected vehicle. In other words, signal information received from the array of single lines of a plurality of sensors (105) will be read in sequence where the detected signal and location variation for a first line sensors (107) of the detector (101) will be recorded as a first reference data as described for Figure 2. All data recorded sequentially after the first line sensors (107) are referenced to the first reference data.

As the vehicle hits the first line sensors (107) from the travel direction of the vehicle, the time and position are recorded and placed as a first point in the chart. Position here refers to the location variation of the array of plurality of sensors (105) arranged in a single line. As it hits the second line sensors (108), the time and the position of the sensors (105) arranged in a single line is recorded as a second point in the chart, until the Nth line sensors in the detector (101) are recorded as described for Figure 2. The data is recorded in the following format:

Ensuing that, the system determines and records subsequent time and position of the vehicle for every succeeding single lines of a plurality of sensors (105), and the subsequent time and position of the detected vehicle are referenced to the initial time and position of the vehicle, wherein the following conditions are made to ensure the detection signal come from the same vehicle, hence any break in the signal detected would mean the end of the detection of the vehicle. The detection for the same vehicle must have adjacent single lines of a plurality of sensors (105) sequence reading regardless if it is ascending or descending. For example, the following sequence of first line sensors (107), second line sensors (108), third line sensors (109), fourth line sensors (110), third line sensors (109), second line sensors (108), and first line sensors (107) is acceptable. However, sequence like of first line sensors (107), second line sensors (108), fourth line sensors (110), second line sensors (108) is not acceptable.

In view of that, each and every single lines of a plurality of sensors (105) must be detected in an unremitting sequential flow; otherwise, when there is a jump in sequence, it will be treated as a new detection for another vehicle.

After which, the system automatically plots a graph of time versus relative position of the vehicle in accordance to the recorded values of the determined initial and subsequent time and position of the detected vehicle, subsequently analyzing the plotted graph and identifying the location and temporal motion pattern of the detected vehicle, and identifying a drive trail derivation of the detected vehicle for determining a driving behavior type.

A reference database to store various driving behavior type is then generated for providing instances of a driving profile assigned to each definition of a driving behavior type. These various driving behavior type stored in the reference database to an assignment to each definition of a driving behavior type using logic, classifier or any combination thereof includes predetermined legal and illegal driving behavior type of driving profile of a U-turn, a right turn, a left turn, forward direction and a sudden stop.

Referring to Figure 4, there is illustrated a process flow of a method for operating the trained system for traffic violation detection in accordance with the present invention (203), wherein operating the system comprises extracting time and position information from a plurality of sensors (105) arranged in a single line, arrayed for a detected vehicle, and processing the time and position information for identifying the location and temporal motion pattern of the detected vehicle, whilst comparing the location and temporal motion pattern information of the detected vehicle with a location and temporal motion pattern of a driving profile created and stored in a reference database during the training of said system for traffic violation detection. This is to provide driving behavior type identification of the detected vehicle. Following that, the system further analyses the identification of said detected vehicle and then proceeds to the step of triggering alarm and capturing image information for a detection of the location and temporal motion pattern of the driving profile matching an illegal driving behavior type. However, the system remains idle for a legal or a normal driving behavior type.

Referring now from Figure 1 to Figure 4, entire course for detecting a vehicle that violates a traffic rule involving the accompanying system is described herein. In the consideration of a case of an illegal U-turn demeanor, information signals detection of a detector (101) at the area of monitor will be transmitted to a processor (103) and processed to identify a drive trail of such illegal U-turn driving behavior type, as well as assigning said behavior type to a profile. Assignment of a driving behavior type profile shall be repeated for various driving behavior types.

o

This driving behavior profile will be then stored away in the storage means (104) as a driving behavior profile database, just before completing the training methods for detecting traffic violation. Next will be the real-time detection of the violating vehicle, where information signals of the violating vehicle detected by the detector (101) are transmitted to the processor

(103) for processing. In the mean time, driving behavior profiles created during the training will be retrieved from the reference database stored in the storage means

(104) to compare the real-time detected information signals of the violating vehicle.

For a matching driving behavior profile compared by means of the processor (103), instructions will be sent from said processor (103) to the capturing means (102) to capture an image of the violating vehicle. Accordingly, the captured image of the violating vehicle will be processed by the processor (103) and stored in the storage means (104) so as to further investigate that violating vehicle. This real-time detection shall be operating constantly as to detect traffic violators. Referring to Figure 5 (a), there is shown a driving behavior type that has been detected by a detector (101) with a corresponding graph representation (201) signifying location variation and time of detected vehicle. As illustrated in the figure, the pattern of the graph provides the behavior information of the driver, where this detection is for a vehicle travelling in straight forward direction.

Referring to Figure 5 (b), there is shown a driving behavior type that has been detected by a detector (101) with a corresponding graph representation (201) signifying location variation and time of detected vehicle. As illustrated in the figure, the pattern of the graph provides the behavior information of the driver, where this detection is for a vehicle making a right turn.

Referring to Figure 6 (a), there is shown a driving behavior type that has been detected by a detector (101) with a corresponding graph representation (201) signifying location variation and time of detected vehicle. As illustrated in the figure, the pattern of the graph provides the behavior information of the driver, where this detection is for a vehicle making a U-turn. As it can be seen that for a U-turn made by the vehicle, the graph merely indicates time sequence taken by a vehicle detected of succeeding single lines of a plurality of sensors (105) for a varying location position of a point that generated along the y-axis of the graph representation (201) which presented extended points in the graph representation (201) in comparison to points generated in previous figures 5 (a) and 5 (b).

Referring to Figure 6 (b), there is shown a driving behavior type that has been detected by a detector (101) with a corresponding graph representation (201) signifying location variation and time of detected vehicle. As illustrated in the figure, the pattern of the graph provides the behavior information of the driver, where this detection is for a vehicle brought to a halt, where the points generated in the graph representation (201) are considerably fewer as compared to previous figures 5 (a), 5 (b), and 6 (a) In as much as the present invention is subject to many variations, modifications and changes in detail, it is intended that all matter contained in the preceding description or shown in the accompanying drawings shall be interpreted as illustrative and not in a limiting sense within the scope of the present invention.

Claims

1. A system for traffic violation detection comprising:

a detector (101) for detecting presence of at least a vehicle; a capturing means (102) for capturing image information of at least a detected vehicle;

a processor (103) for processing information received from the detector (101) and mutually operating with the capturing means (102); and

a storage means (104) mutually operating with the processor (103) for retrieving and storing information of said detector (101), said processor (103), and said capturing means (102);

characterized in that said detector (101) includes a plurality sensors (105) arranged in a single line, arrayed for tracking the presence of the vehicle.

2. A system for traffic violation in accordance to claim 1, wherein said capturing means (102) is preferably a traffic camera to capture snapshots of a violating vehicle.

3. A system for traffic violation in accordance to claim 1, wherein said processor (103) preferably receives digital input from the detector (101) for motion information detection.

4. A system for traffic violation in accordance to claim 1, wherein said processor (103) is preferably for processing information from said detector (101).

5. A system for traffic violation in accordance to claim 1, wherein said processor (103) is preferably for processing information mutually with said capturing means (102).

6. A system for traffic violation in accordance to claim 1, wherein said storage means (104) is preferably a storage device.

7. A system for traffic violation in accordance to claim 1, wherein said plurality of sensors (105) are preferably photosensitive elements.

8. A system for traffic violation in accordance to claim 1, wherein said processor (103) is preferably for processing information mutually with said storage means (104).

9. A method for traffic violation detection comprises:

training a system for traffic violation detection; and

operating the trained system for traffic violation detection.

10. A method for traffic violation detection in accordance to claim 8, wherein training a system for traffic violation detection characterized in that:

receiving a series of timing and location information from a plurality of sensors (105) arranged in a single line, arrayed for creating instances of references to be referred during the training of said system for traffic violation detection;

determining an initial detection of a vehicle for indicating the initial time and position of a vehicle from a first line sensors (107), and recording the initial time and position of the detected vehicle;

determining and recording subsequent time and position of the vehicle for every single lines of a plurality of sensors (105), and the subsequent time and position of the detected vehicle are referenced to the initial time and position of the vehicle;

plotting a graph of time versus relative position of the vehicle in accordance to the recorded values of the determined initial and subsequent time and position of the detected vehicle;

analyzing the plotted graph and identifying a drive trail derivation of the detected vehicle for determining a driving behavior type, and identifying the location and temporal motion pattern of the detected vehicle; and

generating a reference database to store various driving behavior type for providing instances of a driving profile assigned to each definition of a driving behavior type.

11. A method for traffic violation detection in accordance to claim 9, wherein said driving profile includes a predetermined legal and illegal driving behavior type.

12. A method for traffic violation detection in accordance to claim 9, wherein said driving profile includes driving profile of a U-turn.

13. A method for traffic violation detection in accordance to claim 9, wherein said driving profile includes driving profile of a right turn.

14. A method for traffic violation detection in accordance to claim 9, wherein said driving profile includes driving profile of a left turn.

15. A method for traffic violation detection in accordance to claim 9, wherein said driving profile includes driving profile of forward direction.

16. A method for traffic violation detection in accordance to claim 9, wherein said driving profile includes driving profile of a sudden stop.

17. A method for traffic violation detection in accordance to claim 9, wherein said driving profile is preferably assigned to each definition of a driving behavior type using logic, classifier or any combination thereof.

18. A method for traffic violation detection in accordance to claim 8, wherein operating the trained system for traffic violation detection characterized in that:

extracting time and position information from a plurality of sensors (105) arranged in a single line, arrayed for a detected vehicle, and processing the time and position information for identifying the location and temporal motion pattern of the detected vehicle;

comparing the location and temporal motion pattern information of the detected vehicle with a location and temporal motion pattern of a driving profile created during the training of said system for traffic violation detection, for providing driving behavior type identification of the detected vehicle; and triggering alarm and capturing an image information of the vehicle for a driving profile of said vehicle matching an illegal driving behavior type.