PTZ Camera Control via Reinforcement Learning for Sports Event Coverage

The patent applies local quality by making different parts of the field have different capture priorities. The capture priority determination unit assigns higher priorities to regions where subjects are likely to be present (e.g., near goal lines, center circle) and lower priorities to other regions. This allows the system to concentrate camera resources on important areas rather than uniformly covering the entire field, reducing the number of cameras needed while maintaining reliable subject capture.

The patent implements partial action by capturing only the necessary portions of the field rather than the entire field. The image cutting unit generates cropped images from specific high-priority regions identified by the capture priority determination unit. This partial capturing approach eliminates redundant coverage of low-priority areas, reducing camera requirements and system complexity while maintaining sufficient coverage for reliable subject capture.

The system applies local quality by differentiating between high-priority and low-priority regions on the field. The capture priority determination unit identifies specific regions (such as areas near goal lines and the center circle) where subject presence is more likely, and allocates camera resources accordingly. This selective approach allows complete field monitoring reliability to be maintained through strategic camera placement rather than uniform distribution, reducing the total number of cameras required.

The patent implements multi-functionality by enabling each camera to serve multiple purposes: capturing subjects in high-priority regions, providing backup coverage for adjacent areas, and contributing to overall field monitoring. The image cutting unit allows a single camera's footage to be processed into multiple cropped views of different high-priority regions, making each camera more versatile and reducing the total camera quantity needed for comprehensive coverage.

The patent applies the extraction principle by removing unnecessary image regions from the processing pipeline. The image cutting unit extracts only the high-priority regions identified by the capture priority determination unit, discarding redundant low-priority areas. This extraction maintains subject detection precision by ensuring all critical regions are captured and processed, while significantly reducing the total data volume that requires processing, thereby lowering energy consumption.

The system implements partial action by processing only the necessary portions of captured images rather than analyzing entire field views. The image cutting unit generates cropped images focused on high-priority regions where subjects are most likely to be present. This partial processing approach maintains detection precision for critical areas while reducing the overall data processing load and associated energy consumption by eliminating redundant processing of low-priority regions.