TH21989Y - Systems and procedures for detecting traffic light violations using image and/or video processing. - Google Patents

Abstract

The image-based traffic light violation detection system consists of an image receiver, a transmitter, and a display unit. The image receiver processes images and/or video to detect vehicles in the traffic lane, analyze vehicle movement, detect traffic light status, read license plates, and record traffic light violations when they occur. It also stores photographic and video evidence of the violation for use by officials in issuing citations, as well as for supporting documentation in such cases.

Claims (8)

1. The system for detecting traffic light violations using image and/or video processing consists of at least one image receiver (1) which contains at least two image receivers (2): a wide-angle image receiver (6) for taking and/or video images that can display traffic light images and images of vehicles violating traffic lights, and a license plate image receiver (7) for taking and/or video images of vehicle license plates. The image receiver (1) is connected to a transmitter (4) for transmitting signal data via the Internet and/or wirelessly to one or more display units (5). A special feature is that at least one image receiver (1) provides a processing unit (3) which consists of a processing system (8) that processes vehicle movement, traffic light status, and license plate readings from images and/or video received from the image receiver (2) and stores them in temporary memory (9). If a vehicle is found to be violating a traffic light, the image and processing data are stored in a storage unit (10) which records and stores information on traffic light violations. The process for detecting traffic light violations using image and/or video processing involves the following steps: Where the processing section (3) begins by retrieving images and/or video from the wide-angle image sensor (6) (S1), then recording the wide-angle image of the event in memory (9) (S2), analyzing the traffic light signal (S3), if there is vehicle movement while the traffic light signals to stop or prohibit movement in that direction, detecting vehicles in the detection area (11) (S4), if there are vehicles in the detection area (11), detecting the vehicle's movement path and finding and reading the vehicle's license plate (S5), if the vehicle enters the offence area (12) while the traffic light signals to stop or prohibit movement into the designated offence area (12), recording the offence image and data as evidence in the storage unit (10), and sending the image and data to the display section (5) (S6), and if it is found that the vehicle in the image did not violate the traffic light signal, the processing section (3) retrieves the image from the image sensor for further analysis and verification. 2. A system for detecting traffic light violations by processing images and/or video signals under claim 1, where the processing unit (3) processes the detection of vehicles using machine learning algorithms, using a convolutional neural network to process images and/or video signals for traffic light detection. 3. A system for detecting traffic violations by processing images and/or video under claim 1 or 2, where the processing unit (3) processes the vehicle's movement path using appropriate correlation, filter, and tracking techniques to produce a deep sort algorithm. 4. A system for detecting violations of traffic signals by processing images and/or video signals under claim 1 or 2 or 3, where the processing unit (3) processes the status of the traffic signal by processing images from the color intensity and brightness of each lamp. 5. A system for detecting traffic signal violations by processing images and/or video signals under claims 1 or 2 or 3 or 4, where the processing unit (3) will detect and read the characters on Thai license plates using a convolutional neural network algorithm learning system. 6. A system for detecting violations of traffic signals by processing images and/or video under claims 1 or 2 or 3 or 4 or 5, where the processing unit (3) for reading the province name will connect data with the vehicle license plate and the ticketing system. 7. A system for detecting traffic signal violations by processing images and/or video signals under claim 1 or 2 or 3 or 4 or 5, where the processing unit (3) is designed as an embedded system or IoT system. 8. A system for detecting traffic signal violations by processing images and/or video signals under claim 1 or 2 or 3 or 4 or 5 or 6, where the wide-angle image sensor (6) and the license plate image sensor (7) can be combined or separated.