LIDAR Object Recognition Using Contour Reliability for Motion Detection

The patent segments the object detection process into multiple independent verification stages: contour point extraction, coordinate range verification, obstruction status checking, and movement state determination. Each stage processes specific features independently and contributes to the overall reliability score, allowing the system to systematically evaluate multiple criteria rather than relying on a single detection mechanism.

The patent introduces a reliability value parameter that dynamically changes based on multiple verification criteria. The system calculates different reliability scores based on contour point coordinates, obstruction status, and movement characteristics, then uses these parameter variations to determine the final movement state. This transforms the binary classification problem into a multi-parameter evaluation system.

The patent performs preliminary actions by pre-defining coordinate ranges for different object types and pre-establishing verification criteria before actual object detection. The system prepares contour point extraction methods and obstruction detection algorithms in advance, so that during runtime, it only needs to execute predefined verification steps rather than creating complex decision logic from scratch.

The system uses the LIDAR data itself to verify multiple criteria without requiring additional external sensors or complex processing equipment. The same sensor data is reused for contour extraction, coordinate verification, and obstruction detection, allowing the system to perform multiple verification functions using existing resources rather than adding complexity through additional components.

The patent implements partial verification by checking only the necessary contour point coordinates rather than processing all detected points. The system verifies whether minimum and maximum first-axis coordinates fall within predefined ranges, performing selective verification on critical points rather than exhaustive checking of all data, thus reducing processing time while maintaining accuracy.

The system pre-defines coordinate ranges and verification thresholds before actual object detection occurs. By preparing acceptance criteria and boundary values in advance, the system avoids complex real-time calculations and can quickly compare detected coordinates against predetermined ranges, significantly reducing processing time during runtime.