Traffic Object Detection Using Focused ROI Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic object detection systems in autonomous and semi-autonomous vehicles are computationally intensive and often fail to accurately detect objects at high ranges, requiring large data transmission from sensors to neural networks, which is inefficient.
Innovation Solution
The method involves determining focused Regions Of Interest (ROIs) in perception data, scaling them to achieve optimal pixel density, and processing using a neural network-based traffic object detection algorithm to improve detection efficiency and reduce data transmission, incorporating techniques like digital or optical zooming, and blending data from various sources such as maps, fast detectors, and LiDAR data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full perception data from sensor system is transmitted to neural network for processing, then detection accuracy is improved, but data transmission requirements and computational load increase
Solution Approach 1:
The patent extracts only the relevant Regions of Interest (ROIs) from the full perception data before transmission to the neural network. By identifying and isolating specific areas containing traffic objects, the system transmits only necessary data portions, reducing overall data transmission requirements while maintaining detection accuracy for those critical regions.
Solution Approach 2:
The patent segments the perception data into multiple Regions of Interest (ROIs) based on detected traffic objects. Each ROI is processed separately through the neural network, allowing the system to focus computational resources on relevant data segments rather than processing the entire perception data set, thereby reducing computational load while preserving detection accuracy.
2Measurement precision
If neural network processes full perception data, then traffic object detection accuracy is improved, but computational intensity increases
Solution Approach 1:
The system extracts only the necessary Regions of Interest (ROIs) containing traffic objects before neural network processing. By removing irrelevant portions of perception data, the computational intensity required for neural network processing is significantly reduced while the detection accuracy for traffic objects within those ROIs is maintained.
Solution Approach 2:
The perception data is segmented into discrete Regions of Interest (ROIs) that contain traffic objects. The neural network processes each ROI separately rather than the entire perception data set, dividing the computational task into smaller, more manageable segments that reduce overall computational intensity while preserving detection accuracy.
3Measurement precision
If perception data is scaled to achieve optimal pixel density, then detection performance is improved, but processing time increases
Solution Approach 1:
The system scales only the identified Regions of Interest (ROIs) to optimal pixel density rather than scaling the entire perception data set. By applying scaling operations only to relevant segments containing traffic objects, the system achieves optimal detection performance for those regions while minimizing the total processing time required.
Data Source
AI summary
Systems and methods of detecting a traffic object outside of a vehicle and controlling the vehicle. The systems and methods receive perception data from a sensor system included in the vehicle, determine a focused Region Of Interest (ROI) in the perception data, scale the perception data of the at least one focused ROI, process the scaled perception data of the focused ROI using a neural network (NN)-based traffic object detection algorithm to provide traffic object detection data, and control at least one vehicle feature based, in part, on the traffic object detection data.