Excluding Ego Parts from Vehicle Collision Detection Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle collision detection systems waste computing power and generate false warnings due to the analysis of ego parts within the camera's field of view, as these parts do not pose a collision risk and can be mistaken for other objects.
Innovation Solution
A neural network-based collision detection system that identifies and excludes ego parts from the analysis area, using a training dataset with exclusion zones for various types of ego parts, allowing for efficient processing and accurate object detection by generating a 3D geometry of the ego parts and combining it with the vehicle's geometry to determine potential collisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the collision detection system analyzes the entire video feed including ego parts, then comprehensive object detection is performed, but computing power is wasted and false collision warnings are generated
Solution Approach 1:
The patent extracts ego parts from the video feed by identifying them through image recognition and creating exclusion zones around them. The collision detection system then processes only the remaining portions of the video feed outside these exclusion zones, eliminating wasted computation on ego parts while maintaining detection accuracy for actual objects.
2Reliability
If the collision detection system analyzes the entire video feed including ego parts, then no analysis area is excluded, but false collision warnings are triggered due to ego part motion
Solution Approach 1:
The system extracts and isolates ego parts through identification and exclusion zone creation, removing them from the analysis area. This prevents ego part motion from being misinterpreted as collision threats, eliminating false warnings while preserving accurate detection of real objects in the remaining video feed.
3Area of stationary object
If ego parts occupy a significant portion of the field of view, then comprehensive coverage is achieved, but processing efficiency decreases due to analyzing unnecessary areas
Solution Approach 1:
The system maintains comprehensive field of view coverage by keeping the camera positioned to capture the entire scene including ego parts. However, it extracts ego parts through identification and exclusion zone creation, allowing the processing system to efficiently analyze only the necessary portions outside exclusion zones, thus maintaining productivity while preserving full scene awareness.
Solution Approach 2:
The video feed is segmented into exclusion zones (containing ego parts) and non-exclusion zones (containing potential objects of interest). This segmentation allows parallel processing where exclusion zones are identified once and then excluded from continuous analysis, while non-exclusion zones receive focused processing attention, improving overall processing efficiency.
Data Source
AI summary
A vehicle includes a vehicle body having a camera and at least one ego part connection. An ego part is connected to the vehicle body via the ego part connection. A collision detection system is communicatively coupled to the camera and is configured to receive a video feed from the camera. The collision detection system being configured to identify an exclusion region of each frame corresponding to the ego part, perform an object detection on a remainder of each frame, and generate a collision detection warning in response to an object being detected by the object detection system.