LiDAR Vehicle Behavior Detection for Wrong-Way Travel
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle detection systems face challenges in accurately determining vehicle behavior due to environmental disturbances and measurement errors, leading to potential false detections, particularly in distinguishing normal vehicles from wrong-way travelers.
Innovation Solution
A detection device utilizing laser radar-generated point cloud information to calculate vehicle position and movement vectors, employing an inner product comparison with a traffic direction reference vector to enhance accuracy in detecting behaviors such as wrong-way traveling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing is used to calculate vehicle position information, then the detection system can identify vehicle behavior, but the position information accuracy is insufficient due to environmental disturbances
Solution Approach 1:
The patent replaces the optical imaging system (camera) with a laser radar (LiDAR) system that uses laser beams to generate point cloud information. This substitution eliminates the sensitivity to environmental disturbances such as light intensity changes and weather conditions that affect camera-based image processing, thereby improving position information accuracy.
Solution Approach 2:
The patent changes the detection parameter from 2D image coordinates to 3D spatial coordinates using point cloud data. By measuring the actual distance from the detection device to the vehicle in three-dimensional space, the system achieves more accurate position information that is not affected by environmental lighting conditions.
2Reliability
If vehicle direction is determined from feature point displacement, then the system can identify wrong-way traveling vehicles, but false detection occurs when displacement is small
Solution Approach 1:
The patent transitions from 2D image plane analysis to 3D spatial analysis by using point cloud information. This adds the depth dimension (distance from detection device) to the analysis, enabling more accurate calculation of vehicle position and movement vectors, which reduces false detections when displacement is small.
Solution Approach 2:
The patent replaces the 2D image-based feature point tracking method with a 3D point cloud-based method using laser radar. This substitution provides more accurate spatial position information and movement vectors, enabling reliable detection of vehicle direction and behavior even when displacement is minimal.
3Reliability
If only direction of movement is considered for behavior detection, then the system can identify vehicle behavior, but measurement errors lead to false detection
Solution Approach 1:
The patent combines multiple parameters (direction of movement and amount of movement per unit time) into a comprehensive behavior detection method. By calculating the movement vector from position information at different times and comparing it with reference vectors, the system achieves more reliable behavior detection that is not susceptible to measurement errors.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Improves the accuracy of vehicle behavior detection by reducing false positives through robust measurement of vehicle movement and direction, enabling precise identification of traffic violations.
Implementation Method 1
the point cloud information generated on the basis of the reflected light of the laser light
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A detection device, which detects a target behavior that is a vehicle behavior subjected to detection in a preset detection area using a laser radar that generates point cloud information by irradiating the detection area with laser light and receiving reflected light resulting from the irradiation with the laser light, includes an acquisition unit configured to acquire the point cloud information, a vehicle detection unit configured to detect a vehicle on the basis of the point cloud information, and a behavior detection unit configured to detect the target behavior on the basis of a direction of movement and an amount of movement per unit time of the vehicle detected by the vehicle detection unit.