作者: Robert Short , Ronald Driggers , Duke Littlejohn , Mike Scholten
DOI: 10.1364/AO.58.006315
关键词: Computer science 、 Infrared 、 Digital image 、 Orientation (computer vision) 、 Artificial intelligence 、 High dynamic range 、 Target acquisition 、 Digital imaging 、 Optics 、 Image processing 、 Night vision 、 Computer vision 、 Perception 、 Digital image processing
摘要: For the past year, authors have been studying a long-wave infrared (LWIR) sensor design concept that combines high detector well capacity, small-pitch detectors, and digital image processing to optimize target acquisition. Theoretical performance modeling [via Night Vision Integrated Performance Model (NVIPM)] suggests our approach offers large increase in identification range, but multiple field trials using triangle orientation discrimination (TOD) yielded results are inconsistent with model's predictions. this reason, we performed human perception experiments on simulated TOD targets, without processing, assess utility of value as an evaluation for enhancement. The do not agree NVIPM range improvement. We present results, along experiment methodologies, guide future investigations into effects boost or restoration filtering performance.