Cross-Modality Vehicle Object Detection With Two-Phase Sensor Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern vehicles, particularly autonomous and semi-autonomous vehicles, face challenges in accurate object detection due to the computational intensity of high-resolution sensors like cameras and Lidar, which limits their ability to process data in real-time without sacrificing accuracy, especially when detecting distant or small objects.
Innovation Solution
A two-phase processing system using cross-modality sensors, where data from a Lidar sensor and a camera are down-sampled to lower resolution for initial rapid processing, followed by higher resolution processing of unconfirmed proposals to verify object identities, reducing computational intensity and power consumption while maintaining accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution sensor data is processed to maintain accurate object detection, then detection accuracy is improved, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent divides object detection into two phases: a glance phase that processes down-sampled data for rapid candidate identification, and a focus phase that processes full-resolution data only for confirmed candidates. This segmentation allows the system to maintain high detection accuracy while significantly reducing overall computational complexity by limiting intensive processing to only necessary cases.
Solution Approach 2:
The system performs preliminary down-sampling and processing of sensor data before full-resolution analysis. By pre-processing data at lower resolution to identify candidate objects, the system prepares the data in advance for more efficient high-resolution processing, reducing the total computational burden while maintaining accuracy.
2Reliability
If full-resolution sensor data is processed to detect distant or small objects, then detection reliability is improved, but processing speed decreases
Solution Approach 1:
The patent segments the detection process into rapid low-resolution screening followed by targeted high-resolution verification. This allows the system to maintain high processing speed through efficient down-sampled data analysis while ensuring detection reliability through selective full-resolution processing of candidate objects.
Solution Approach 2:
The system applies partial high-resolution processing only to candidate objects identified in the glance phase, rather than processing all data at full resolution. This partial action approach maintains detection reliability for critical objects while significantly improving overall processing speed.
3Measurement precision
If high-resolution data from multiple sensors is processed simultaneously, then object identification accuracy is improved, but power consumption increases
Solution Approach 1:
The patent segments sensor data processing into two resolution levels, processing down-sampled data from multiple sensors during the glance phase and only processing full-resolution data for confirmed candidates during the focus phase. This segmentation maintains accurate object identification while significantly reducing power consumption by avoiding simultaneous high-resolution processing of all sensor data.
Solution Approach 2:
The system performs preliminary processing of multi-sensor data at reduced resolution before committing to high-resolution analysis. This preliminary action allows the system to evaluate multiple sensor inputs efficiently and only intensively process data when necessary, reducing overall power consumption while maintaining identification accuracy.
Data Source
AI summary
A system includes first and second sensors and a controller. The first sensor is of a first type and is configured to sense objects around a vehicle and to capture first data about the objects in a frame. The second sensor is of a second type and is configured to sense the objects around the vehicle and to capture second data about the objects in the frame. The controller is configured to down-sample the first and second data to generate down-sampled first and second data having a lower resolution than the first and second data. The controller is configured to identify a first set of the objects by processing the down-sampled first and second data having the lower resolution. The controller is configured to identify a second set of the objects by selectively processing the first and second data from the frame.