Optical Sensor Region Segmentation for Autonomous Vehicle Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional self-driving car object detection techniques require excessive calculation resources and may compromise real-time processing and power efficiency, especially when detecting pedestrians, as they often perform detection processing multiple times, including after vehicle detection, which increases resource usage and does not effectively improve detection performance in regions without vehicles.
Innovation Solution
An information processing device that obtains sensing data from optical sensors, combines it with position and map information to determine specific sensing data regions, and uses this information to input relevant data to an object detection model, optimizing processing by narrowing the target and allocating resources based on the specific sensing data regions, such as sidewalks, to improve detection performance while reducing resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detection processing is performed multiple times (e.g., vehicle detection followed by pedestrian detection), then object detection performance is improved, but calculation resources and processing time increase significantly
Solution Approach 1:
The patent divides the sensing region into multiple regions of interest (ROIs) based on map information (e.g., sidewalks, roads, parks). Instead of performing detection processing uniformly across the entire sensing region, the system segments the area and applies detection only to relevant ROIs. This segmentation approach maintains detection performance in critical areas while reducing overall calculation resources required.
Solution Approach 2:
The patent applies different detection strategies to different regions based on their characteristics. High-resolution detection is applied to regions where pedestrians are likely to appear (e.g., sidewalks near vehicles), while lower-resolution or no detection is applied to regions where pedestrians are unlikely to appear. This local quality approach optimizes the balance between detection performance and resource consumption.
2Reliability
If detection processing is performed multiple times, then detection coverage is improved, but processing time increases and real-time performance deteriorates
Solution Approach 1:
The patent performs preliminary classification of the sensing region using map information before conducting detailed object detection. By pre-identifying regions of interest based on geographic data (sidewalks, roads, parks), the system prepares the detection framework in advance, allowing subsequent detection to focus only on relevant areas. This preliminary action reduces the time required for comprehensive detection while maintaining coverage where needed.
Solution Approach 2:
The patent dynamically adjusts the detection strategy based on the specific scene context provided by map information. The system adapts which regions require detection and at what resolution, rather than applying a static uniform detection approach. This dynamic adaptation allows the system to maintain high detection coverage in critical areas while reducing processing time in less critical areas.
3Reliability
If uniform detection processing is applied to the entire sensing region, then detection coverage is maximized, but power consumption and calculation resources increase
Solution Approach 1:
The patent extracts and processes only the essential information needed for detection by utilizing pre-existing map data. Instead of analyzing every pixel or data point in the entire sensing region, the system extracts relevant regions of interest based on map information and focuses computational resources only on those extracted areas. This extraction approach maintains detection coverage while significantly reducing power consumption.
Solution Approach 2:
The patent changes the detection parameters (such as resolution, detection threshold, and processing intensity) based on the specific region being analyzed. Regions identified as high-priority through map information receive higher detection parameters, while other regions receive lower parameters or are excluded from detection. This parameter adaptation allows the system to maintain adequate detection coverage across the scene while optimizing power consumption by avoiding uniform high-resource processing everywhere.
Data Source
AI summary
An information processing device is configured to: obtain sensing data from an optical sensor; obtain position information of a mobile body which includes the optical sensor; obtain map information; determine, in sensing data, a specific sensing data region corresponding to a specific region in a sensing region of the optical sensor by using the position information and the map information that have been obtained; determine input information to be provided to an object detection model, according to the specific sensing data region; and cause the object detection model to perform object detection processing by using the input information.