Multi-View Human Detection via Rectified Coordinate Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current factory automation systems lack the ability to naturally and efficiently cooperate with humans, as they do not effectively track human position and behavior in dynamic environments, leading to potential interactions with automated equipment.
Innovation Solution
A vision monitoring system comprising multiple cameras and a visual processor that captures time-synchronized video feeds, detects human presence and pose using pattern matching algorithms, and alters the behavior of automated equipment to prevent interactions by fusing image frames into a common coordinate system and tracking human motion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple cameras are used to capture video feeds from different viewpoints, then the ability to detect humans in dynamic environments is improved, but the system complexity and computational load increase
Solution Approach 1:
The system divides the workspace into multiple regions of interest (ROIs) and processes image frames from different cameras independently for each ROI. This segmentation allows parallel processing of multiple video feeds, reducing the computational burden on a single processor while maintaining comprehensive coverage of the workspace area.
Solution Approach 2:
The system transforms 3D spatial information from multiple camera viewpoints into a unified 2D ground plane representation. By projecting detections from multiple cameras onto a common 2D workspace map, the system integrates multi-view data without requiring complex 3D reconstruction, thus reducing computational complexity while preserving detection accuracy.
2Measurement precision
If pattern matching is performed on the entire image frame, then detection coverage is improved, but processing time increases
Solution Approach 1:
The system extracts only the relevant region of interest (ROI) from each image frame based on the camera's field of view and the workspace layout. By applying pattern matching exclusively to these extracted ROIs rather than entire image frames, the system maintains comprehensive detection coverage within the workspace while significantly reducing the number of pixels processed, thus decreasing processing time.
3Reliability
If the system tracks human position and behavior in real-time, then safety is improved, but computational resources are consumed
Solution Approach 1:
The system performs preliminary processing by detecting human presence and extracting key features (such as body orientation and motion patterns) before full behavioral analysis. This preliminary detection stage allows the system to maintain real-time safety monitoring by quickly identifying humans and their basic states, while more detailed behavioral analysis is performed only when necessary, thus reducing overall computational resource consumption.
Data Source
AI summary
A human monitoring system includes a plurality of cameras and a visual processor. The plurality of cameras are disposed about a workspace area, where each camera is configured to capture a video feed that includes a plurality of image frames, and the plurality of image frames are time-synchronized between the respective cameras. The visual processor is configured to receive the plurality of image frames from the plurality of vision-based imaging devices and detect the presence of a human from at least one of the plurality of image frames using pattern matching performed on an input image. The input image to the pattern matching is a sliding window portion of the image frame that is aligned with a rectified coordinate system such that a vertical axis in the workspace area is aligned with a vertical axis of the input image.