Driver Camera Occupancy Tracking via Foreground Pixel Counting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driver-facing camera monitoring systems inefficiently use computing resources due to data-intensive video streams, often analyzing only video data around specific events like accidents, limiting effective monitoring of driver behavior in other situations.
Innovation Solution
A method that captures video image data from a driver-facing camera, defines areas-of-interest, determines foreground pixel counts, maintains occupancy history, and initiates external indicators based on final status determinations, using frame-to-frame differencing and edge detection techniques to efficiently monitor driver presence and position.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If video streams are analyzed in real-time to monitor driver behavior continuously, then monitoring effectiveness is improved, but computing resource consumption increases
Solution Approach 1:
The patent divides the video stream analysis into two segments: real-time analysis of only area-of-interest regions (driver's face and upper body) using simplified algorithms, and comprehensive analysis of selected video clips using full video processing. This segmentation allows continuous monitoring with reduced computational load while maintaining reliability through periodic full analysis of critical events.
Solution Approach 2:
The system performs partial analysis by focusing computational resources only on relevant portions of the video stream (area-of-interest) rather than processing the entire video frame continuously. This partial action approach maintains monitoring effectiveness for critical regions while significantly reducing overall computing resource consumption.
2Use of energy by moving object
If only video data around specific events is analyzed, then computing resources are conserved, but driver behavior monitoring is limited to accident situations only
Solution Approach 1:
The system performs preliminary real-time analysis of area-of-interest regions throughout the video stream to detect potential driver behavior issues before they result in accidents. This preliminary action enables early detection and continuous monitoring of driver state (alertness, distraction, fatigue) while conserving resources by avoiding full video analysis during normal driving conditions.
Solution Approach 2:
The system implements periodic full video analysis at scheduled intervals or triggered by specific conditions, complementing continuous lightweight area-of-interest monitoring. This periodic comprehensive analysis ensures broad monitoring coverage and captures events that may occur outside predefined areas, maintaining reliability while managing computational resources efficiently.
Data Source
AI summary
A method for determining a presence of an occupant within a vehicle includes capturing, via a driver facing camera, video image data, including a plurality of image frames of a field-of-view of the driver facing camera. At least one area-of-interest within the image frames is defined, and a foreground pixel count of each image frame of each area-of-interest is determined. At each Nth image frame, an occupancy for each area of interest is determined based at least in part on the foreground pixel count. The occupancy indicates whether the occupant is present or absent in the respective area-of-interest. A history of the occupancy determination for each area-of-interest is maintained, and a final status for each area-of-interest is determined based on the respective history of occupancy. External indicators are initiated based on the final status determination.