计算机科学
人工智能
旋光法
计算机视觉
模式识别(心理学)
遥感
目标检测
自动目标识别
杂乱
探测器
特征提取
变更检测
恒虚警率
作者
Cornell S. L. Chun,Firooz A. Sadjadi
摘要
Abstract : When thermal imaging sensors are used for automatic target detection and recognition of ground targets, the performance is often limited by the small thermal contrast between target and background. In comparison, polaimetric imaging is minimally dependent on the thermal characteristics of the target. Polarimetric data depends, instead, on the three-dimensional shape and surface properties of the target. Physics Innovations Inc. and Lockheed Martin Corp, have an ongoing program to develop polarimetric imaging systems for automatic target detection and recognition. In this paper, we will discuss two novel technologies which we have recently developed for these imaging systems. The first is a two-dimensional array of microscale retarders. This retarder array, when combined with a two-dimensional array of linear polarizers and a focal plane array, will enable the imaging of the four Stokes parameters at video frame rates without the mechanical rotation of optical components. The retarder elements are highly achromatic. Furthermore, the retarders have a large field angle. A large field angle will facilitate the use of low f-number optics for greater sensitivity. The second technology developed for polaimetric imaging systems are target detection and recognition algorithms which optimally combine the Stokes images.
科研通智能强力驱动
Strongly Powered by AbleSci AI