ADAS Object Detection Using Stereo-Radar Bounding Box Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Advanced Driver Assist Systems (ADAS) face challenges in effectively detecting objects using signals from multiple sensors, leading to inefficiencies in identifying relevant bounding boxes and determining vehicle risks due to external factors.
Innovation Solution
The ADAS employs a processing circuit with a position information generation engine, tracking list generation engine, object detector, and object tracking engine to generate and adjust class scores of candidate bounding boxes based on stereo images and reflected signals from cameras and radar, selecting the most relevant bounding box for object detection and risk assessment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the ADAS uses signals from multiple sensors to detect objects, then the detection coverage is improved, but the complexity of processing and identifying relevant bounding boxes increases
Solution Approach 1:
The patent combines bounding box information from multiple sensors (camera and radar) by merging candidate bounding boxes based on their spatial overlap and confidence scores. This integration approach consolidates data from different sensor sources into a unified object detection result, improving reliability while managing processing complexity through systematic combination rules.
Solution Approach 2:
The patent introduces an intermediary processing stage that generates candidate bounding boxes from sensor signals and then selectively refines them based on confidence thresholds and spatial relationships. This intermediary layer acts as a mediator between raw sensor data and final object detection, filtering and prioritizing information to reduce processing complexity.
2Measurement precision
If the ADAS generates multiple candidate bounding boxes for object detection, then the detection accuracy is improved, but the time required to process and select relevant bounding boxes increases
Solution Approach 1:
The patent performs preliminary actions by pre-calculating confidence scores and spatial relationships for candidate bounding boxes before final selection. By preparing and organizing bounding box data in advance with associated metadata (confidence levels, sensor sources, spatial coordinates), the system reduces the time required for final selection and risk assessment.
Solution Approach 2:
The patent dynamically adjusts processing parameters such as confidence thresholds and selection criteria based on driving conditions and risk levels. By changing parameters adaptively, the system can process fewer high-confidence candidates in low-risk situations while maintaining high detection accuracy when needed, thus reducing average processing time.
3Measurement precision
If the ADAS adjusts class scores of candidate bounding boxes based on position information matching, then the detection accuracy is improved, but the computational load increases
Solution Approach 1:
The patent applies partial action by selectively adjusting class scores only for candidate bounding boxes that meet specific criteria (spatial overlap threshold, confidence level threshold). Instead of processing all candidate boxes uniformly, the system focuses computational energy on promising candidates, reducing overall computational load while maintaining detection accuracy for relevant objects.
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 and efficiency of object detection by reducing the number of non-selected candidate bounding boxes, improving the system's ability to determine vehicle risks and provide timely notifications.
Implementation Method 1
generating a second position information associated with the at least one object based on reflected signals received from the vehicle that is in motion
Data Source
AI summary
ADAS includes a processing circuit and a memory which stores instructions executable by the processing circuit. The processing circuit executes the instructions to cause the ADAS to receive, from a vehicle that is in motion, a video sequence, generate a position image including at least one object included in the stereo image, generate a second position information associated with the at least one object based on reflected signals received from the vehicle, determine regions each including at least a portion of the at least one object as candidate bounding boxes based on the stereo image and the position image, and selectively adjusting class scores of respective ones of the candidate bounding boxes associated with the at least one object based on whether a respective first position information of the respective ones of the candidate bounding boxes matches the second position information.