Safe Area Monitoring With Refined Obstacle Boundary Detection

The patent segments the obstacle detection process into two distinct stages: first, object detection technology identifies the obstacle and generates an initial obstacle area; second, contour recognition technology processes the obstacle boundary to generate a refined obstacle contour. This segmentation allows each technology to operate optimally in its specialized domain, with contour recognition correcting the boundary precision issues of object detection.

The patent introduces contour recognition technology as an intermediary processing step between object detection and collision determination. This intermediary refines the rough obstacle boundary from object detection into a precise obstacle contour that accurately represents the actual obstacle boundary, thereby resolving the measurement precision problem without sacrificing detection capability.

The patent divides the collision determination process into two sequential stages: first, determine whether the obstacle area intersects with the safe area; second, determine whether the refined obstacle contour intersects with the safe area. This segmentation ensures that both the broader detection coverage and the precise boundary information are utilized, improving reliability by reducing false positives while maintaining comprehensive detection.

The patent implements a feedback mechanism where the results from obstacle area intersection detection feed into the more precise obstacle contour intersection detection. This two-stage feedback process allows the system to efficiently filter out obvious cases while applying more computationally intensive contour analysis only when necessary, improving overall reliability without excessive computational burden.

The patent applies preliminary action by first performing the faster object detection to identify obstacles and generate initial obstacle areas. This preliminary step quickly screens the environment for potential obstacles, and only when obstacles are detected does the system proceed to the more time-consuming contour recognition process. This preliminary action maintains processing speed while ensuring precision when needed.

The patent segments the processing pipeline into speed-optimized and precision-optimized stages. Object detection technology provides rapid initial identification, while contour recognition technology provides precise boundary determination. This segmentation allows the system to maintain high processing speed for most scenarios while achieving high measurement precision when obstacles are present.