论文部分内容阅读
针对传统的像素级遥感图像变化检测方法中检测精度严重依赖于图像配准、辐射校正和差异图像阈值选取的问题,提出了一种基于目标匹配的目标级遥感图像变化检测方法。该方法直接作用于2幅未经配准和辐射校正的不同时相遥感图像,利用目标的区域不变矩与目标之间的角度关系对目标进行匹配,将不能匹配的目标作为变化目标,利用匹配目标对2幅图进行配准,以同时实现图像的配准与变化检测,并用配准结果对变化检测结果进行修正。实验结果表明,该方法具有较好的定性检测性能。
Aiming at the problem that the detection accuracy of the traditional pixel-level remote sensing image change detection method relies heavily on image registration, radiation correction and difference image threshold selection, a change detection method of target-level remote sensing image based on target matching is proposed. The method works directly on two remote sensing images without registration and radiation correction in different time series. The target is matched by using the angular relationship between the target’s moment invariants and the target, and the target that can not be matched is taken as the target of change Matching target to register two images to achieve the registration and change detection of the image at the same time, and use the registration result to correct the change detection result. Experimental results show that this method has good qualitative detection performance.