Vehicle Speed Estimation Using 3D Polyhedrons Without Camera Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera-based speed estimation systems require calibration for accurate speed measurement, which is time-consuming and prone to errors due to environmental changes, and conventional 2D detections are less precise.
Innovation Solution
An image processing method using uncalibrated cameras that derives polyhedrons from vehicle images to determine vehicle size information, allowing for speed estimation without calibration, by assuming known vehicle sizes and geometric properties to calculate distance and time differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If calibration is used for camera-based speed estimation, then measurement precision is improved, but device complexity and time consumption increase
Solution Approach 1:
The patent extracts the calibration requirement from the speed estimation system by using vehicle size information as a reference object. Instead of calibrating the camera system itself, the method uses known vehicle dimensions to establish the relationship between image pixels and real-world distances, thereby eliminating the need for complex camera calibration while maintaining measurement accuracy.
Solution Approach 2:
The patent introduces vehicle size information as an intermediary element to bridge the gap between image detection and real-world measurement. By using the known dimensions of vehicles as a reference, the system can calculate distances and speeds without direct camera calibration, thus simplifying the overall system while preserving measurement precision.
2Device complexity
If 2D detection is used for vehicle detection, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent transitions from 2D detection to 3D polyhedron representation to capture the spatial structure of vehicles more accurately. By representing vehicles as three-dimensional polyhedrons with defined vertices and faces, the system gains additional spatial information that improves detection precision while maintaining relatively simple processing requirements.
Solution Approach 2:
The patent segments the vehicle detection process into multiple components: detecting vehicle contours, determining polyhedron vertices, calculating polyhedron dimensions, and deriving speed information. This segmentation allows each component to be optimized independently, improving overall precision without proportionally increasing system complexity.
Data Source
AI summary
An image processing method for determining a travel speed of a vehicle. A monitoring scene with at least one vehicle is arranged in the monitoring region, where at least a first image of the monitoring scene is detected. A first time stamp is assigned to the first image, and a second image of the monitoring scene is detected, a second time stamp is assigned to the second image, and a first polyhedron is derived from the first image and a second polyhedron is derived from the second image. The first and the second polyhedron each represent the vehicle. Vehicle size information is determined for the first polyhedron and/or for the second polyhedron, A distance travelled by the vehicle in the monitoring region is established based on the vehicle size information and the polyhedrons, and the vehicle speed is established based on the distance and the time stamp.