Two-Wheeler Detection Region Expansion for Radar-Camera Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The detection accuracy of two-wheelers using a monocular camera is poor, leading to potential erroneous identification as non-identical objects when combined with radar data, due to large errors in detected positions.
Innovation Solution
The object detection apparatus enlarges the image search region when determining an object as a two-wheeler, ensuring overlap with the radar search region to accurately identify two-wheelers using both radar and monocular camera data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the monocular camera is used to detect object position, then the cost is reduced compared to stereo camera, but the detection accuracy for distance direction deteriorates
Solution Approach 1:
The patent combines radar detection results with monocular camera detection results through association processing. The radar provides accurate distance information while the camera provides visual identification, merging their complementary strengths to achieve both cost reduction and maintained detection accuracy.
Solution Approach 2:
The patent introduces an association processing unit as an intermediary that correlates radar detection data with camera detection data. This mediator resolves the accuracy issue by using radar's strong distance measurement capability to compensate for the monocular camera's weakness in depth perception.
2Measurement precision
If the image search region is enlarged to accommodate two-wheeler detection errors, then the detection accuracy of two-wheelers is improved, but the computational load and processing time increase
Solution Approach 1:
The patent applies different search region sizes based on object type. For two-wheelers, a larger image search region is used to accommodate detection errors, while for other objects, the standard region size is maintained. This localized adaptation improves two-wheeler detection without unnecessarily increasing processing time for all objects.
Solution Approach 2:
The patent dynamically adjusts the image search region size based on the detected object type. When a two-wheeler is detected, the system automatically enlarges the search region; otherwise, it uses the standard region. This dynamic adjustment optimizes the balance between detection accuracy and processing efficiency.
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
This approach enhances the accuracy of two-wheeler detection by compensating for the large errors in monocular camera positioning, ensuring precise identification and reducing false detection rates.
Implementation Method 1
an object detection apparatus detecting an object existing around an own vehicle, using a radar and a monocular camera
Implementation Method 2
a monocular camera detecting an object existing around an own vehicle
Data Source
AI summary
An object detection apparatus includes a radar-region-setting-unit setting a detection region of an object as a radar search region based on a detected position of the object detected by a radar, an image-region-setting-unit setting the detection region of the object as an image search region based on the detected position of the object detected by a monocular camera, an object-determination-unit determining the object detected by a radar and the monocular camera as the same object under a condition that an overlapping portion of the radar search region and the image search region exists, and a two-wheeler-determination-unit determining whether an object type detected by the monocular camera is a two-wheeler. The image-region-setting-unit enlarges the image search region compared with a case where the object type is not the two-wheeler, under a condition that the object type is determined as the two-wheeler.