Vision-Based Red Light Camera False Positive Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Red Light Camera Systems (RLCS) generate a high number of false detections, leading to erroneous tickets, as they fail to accurately distinguish between vehicles that run a red light and those that abruptly stop, requiring time-consuming manual review by law enforcement officials.
Innovation Solution
A vision-based system and method that processes evidentiary images by extracting feature points, computing feature descriptors, determining correspondences, and extracting attributes to automatically differentiate between vehicles that run a stop signal and those that abruptly stop, using a processor configured to acquire and analyze images from cameras monitoring intersections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual review by law enforcement officials is used to determine violations, then accuracy in distinguishing violating vehicles from non-violating vehicles is improved, but time consumption and processing efficiency deteriorate
Solution Approach 1:
The patent replaces the mechanical system of manual review by law enforcement officials with an automated image processing system that uses feature extraction, descriptor computation, and correspondence matching to automatically analyze evidentiary images and determine whether vehicles violated traffic signals, thereby eliminating time-consuming manual review while maintaining detection accuracy
Solution Approach 2:
The system enables self-service by allowing the image processing system to automatically perform the violation determination task that previously required human officials, using algorithms to extract features, match correspondences across multiple images, and autonomously classify vehicles as violating or non-violating without human intervention
2Productivity
If existing RLCS automatically detect vehicles using sensor triggers, then processing speed and automation are improved, but false detection rate increases
Solution Approach 1:
The patent applies segmentation by dividing the violation detection task into multiple stages: acquiring multiple evidentiary images, extracting features from each image, computing descriptors, determining correspondences across images, and finally classifying the vehicle as violating or non-violating based on the aggregated evidence, which reduces false detections by analyzing multiple aspects of vehicle behavior rather than relying on a single sensor trigger
Solution Approach 2:
The system uses feedback by comparing vehicle position and motion across multiple evidentiary images captured at different time points, using the extracted features and correspondences to determine whether the vehicle actually violated the traffic signal or merely stopped abruptly, thereby providing feedback validation that reduces false detections while maintaining automated processing speed
3Reliability
If multiple evidentiary images are processed with feature extraction and correspondence matching, then false positive reduction is improved, but computational complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing feature descriptors for each evidentiary image before performing correspondence matching, and by pre-defining the classification rules for determining violations, which organizes the computational complexity into manageable preprocessing and processing stages, reducing the false positive rate through systematic analysis
Solution Approach 2:
The system uses parameter changes by transforming the raw image data into feature space through extraction and descriptor computation, then using correspondence parameters to determine vehicle behavior, which changes the parameter representation from pixel values to meaningful feature descriptors, enabling more reliable violation detection despite increased processing complexity
Data Source
AI summary
A method for detecting a vehicle running a stop signal includes acquiring at least two evidentiary images of a candidate violating vehicle captured from at least one camera monitoring an intersection. The method includes extracting feature points in each of the at least two evidentiary images. The method includes computing feature descriptors for each of the extracted feature points. The method includes determining a correspondence between feature points having matching feature descriptors at different locations in the at least two evidentiary images. The method includes extracting at least one attribute for each correspondence. The method includes determining if the candidate violating vehicle is in violation of running the stop signal using the extracted attribute.