遮挡物体的图像恢复方法研究与仿真

发布时间：2018-05-03 18:36

本文选题：遮挡物体 + 图像恢复　；参考：《计算机仿真》2016年11期

【摘要】：针对传统的图像恢复方法,只能对背景简单且没有遮挡物的目标图像进行恢复,对识别目标背景复杂且存在物体遮挡的图像不能准确进行恢复,导致恢复性能差的问题。提出基于阻尼最小二乘法的遮挡物体的图像恢复方法。对原始遮挡物图像存在的大量噪声,采用回溯双边滤波法对原始图像进行去噪处理,依据阈值法将预处理后的图像划分成特定数量的具有一定意义的小区域,并寻找其中的兴趣目标区域,根据阻尼最小二乘法理论,对被遮挡物体的轮廓进行复原,实现遮挡物体的图像恢复。仿真结果表明,采用改进的算法相比传统的方法,均方误差和平均绝对误差均有所降低,峰值信噪比得到提高,具有较强的抗噪能力,能够满足实际的应用需求。
[Abstract]:The traditional image restoration method can only restore the target image with simple background and no occlusion, but can not recover the target image with complex background and object occlusion accurately, which leads to the problem of poor recovery performance. A method for image restoration of occluded objects based on damped least square method is proposed. For the large amount of noise existing in the original occlusion image, the backtracking bilateral filtering method is used to de-noise the original image, and the pre-processed image is divided into a certain number of small areas with certain significance according to the threshold method. Based on the theory of damped least square method, the contour of the occluded object is restored and the image of the occluded object is restored. The simulation results show that compared with the traditional method, the improved algorithm can reduce the mean square error and average absolute error, improve the peak signal-to-noise ratio, and have a strong anti-noise ability, which can meet the practical application needs.
【作者单位】：河南师范大学新联学院;
【分类号】：TP391.41

【相似文献】