Vehicle Target Detection Y-Axis Accuracy via Bearing and Width
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing target detection systems for vehicles face accuracy issues in determining the Y-coordinate of targets, leading to incorrect recognition of single targets as multiple targets, especially as the distance from the vehicle increases, requiring an excessively large error region for fusion determination.
Innovation Solution
A target detection device that includes a radar specification unit, an image specification unit, and a same target determination unit, which sets the Y-axis range of the image detection region based on the bearing from the vehicle to the target's ends and expected width values, ensuring accurate Y-coordinate calculation and reducing the likelihood of incorrect target recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the Y-coordinate is detected by detecting the height of the lower end position of the target in the captured image, then the detection method is simple, but the accuracy of detecting the Y-coordinate is reduced as the distance to the target increases
Solution Approach 1:
The patent changes the parameter used for Y-coordinate detection from the height of the lower end position to the X-directional width of the detection target. By using width information instead of height information, the system achieves more accurate distance-based Y-coordinate detection without increasing system complexity
Solution Approach 2:
The patent introduces an intermediary calculation process that uses the detected X-directional width as a mediator to derive the Y-coordinate. Instead of directly using height, the system calculates width first and then uses it to determine the Y-coordinate through geometric relationships
2Reliability
If an excessively large error region is set to prevent incorrect recognition, then fusion determination reliability improves, but the image detection region becomes unnecessarily large
Solution Approach 1:
The patent changes the parameter for defining the detection region from a fixed large error region to a dynamically calculated Y-axis range based on bearing and expected width values. This allows the detection region to be precisely sized according to actual target characteristics rather than using an excessively large safety margin
Solution Approach 2:
The patent makes the detection region dynamic by calculating the Y-axis range based on the bearing to the target and expected width values. The region adapts to different target distances and sizes, maintaining optimal detection coverage without unnecessary expansion
3Device complexity
If the Y-axis range is not appropriately set, then processing is simpler, but target recognition accuracy decreases
Solution Approach 1:
The patent performs preliminary calculation of the Y-axis range using bearing and expected width values before target recognition. By pre-establishing the appropriate detection region boundaries, the system ensures accurate target recognition without adding complex real-time processing requirements
Data Source
AI summary
A target detection device mounted in a vehicle includes: a radar specification unit; a same target determination unit; and an image specification unit that specifies, as image target position information, an image detection region of an image detection target detected on the basis of a captured image, the image detection region representing the position of the image detection target relative to the base point on an XY-plane with the X-axis representing the vehicle width direction of the vehicle and the Y-axis representing the vehicle length direction of the vehicle. The image specification unit sets the Y-axis range of the image detection region in the Y-axis direction on the basis of the bearing from the base point to each of the two X-directional ends of the image detection target in the captured image and of an expected width minimum value and an expected width maximum value in the X-axis direction.