论文部分内容阅读
提出了一种基于多尺度背景纹理分析的目标检测方法．通过小波多尺度分解来提取背景纹理的能量特征，并计算这些特征向量与中心向量之间的距离，目标检测在所得的距离像上完成．利用红外小目标的特性，根据距离像统计直方图来实现门限的自适应选取．实验验证了该方法的有效性
A target detection method based on multi-scale background texture analysis is proposed. The energy features of the background texture are extracted by wavelet multi-scale decomposition, and the distance between these feature vectors and the center vector is calculated. The target detection is completed on the resulting distance image. Using the characteristics of infrared small target, the threshold is adaptively selected according to the distance like statistical histogram. Experiments show the effectiveness of the method