一种改进形态学建筑物指数的无人机影像变化检测方法

A Method for UAV Image Change Detection Based on Improved Morphological Building Index

无人机遥感影像具有配准精度高、分辨率高的特点,针对无人机影像成像特征,提出了一种改进BMI的无人机影像建筑物检测方法。该法顾及检测区域纹理、光谱和形态特征,通过改进其自适应对比度和结构元素的尺度特征,扩大建筑物与背景波谱特征,实现建筑物变化检测。基于赣州某区两时相无人机遥感影像,进行了基于改进BMI的无人机变化检测,结果发现改进方法具有较高的检测精度和抗干扰性,与MBI方法相比,漏检率下降6.55%,错检率下降5.08%,总体精度提高了7.89%,不失为一种高效的无人机影像建筑物变化检测新方法。

UAV remote sensing images have the characteristics of high registration accuracy and resolution.Aiming at the characteristics of UAV imaging,an improved BMI building detection method from UAV images is proposed.Considering the texture,spectral,and morphological features of the detection region,the method improves the adaptive contrast and the scale of the structure elements,enlarges the spectral characteristics between building and background,and thus to realize the building change detection.Taking two periods of UAV images in a region of Ganzhou,based on improved BMI,we carried on the UAV image change detection experiment.AS compared with the MBI method,the proposed method improves the precision and anti-jamming,the omission rate is decreased by 6.55%,commission rate is decreased by 5.08%,and the overall accuracy is increased by 7.89%.It proves to be a new efficient building change detection method for UAV images.