不同模糊尺度下运动模糊方向的鉴别

发布时间：2018-11-02 19:27

【摘要】：为了有效减弱高动态环境对图像采集的干扰,利用模糊图像的频谱特性,提出了一种适合不同模糊尺度的运动模糊方向估计方法。该方法对比不同模糊尺度图像频谱中心亮条纹边缘特征的差别,通过计算其曲率半径实现不同模糊尺度图像的量化鉴别,在确定图像模糊尺度的大小之后,分别利用倒频谱、二次傅氏变换对模糊图像频谱中心亮线特征进行提取,考虑图像高宽比、中心十字亮线的不利影响,在二值化处理的基础上剔除十字亮线,通过亮线特征点直线拟合完成模糊方向估计。实验结果表明,该算法能够有效鉴别不同模糊尺度的运动模糊方向,估计精度较其他同类算法更高,相对误差率均小于0.02。
[Abstract]:In order to effectively reduce the interference of high dynamic environment to image acquisition, a motion blur direction estimation method suitable for different fuzzy scales is proposed by using the spectral characteristics of blurred images. In this method, the difference of the edge features of the center bright stripes of different fuzzy scale images is compared, and the quantitative identification of different fuzzy scale images is realized by calculating its curvature radius. After the size of the fuzzy scale of the image is determined, the cepstrum is used respectively. The second Fourier transform is used to extract the feature of the center bright line of the blurred image spectrum. Considering the adverse effects of the aspect ratio of the image and the center cross bright line, the cross bright line is eliminated on the basis of the binary processing. The fuzzy direction estimation is accomplished by fitting the feature points of bright lines. Experimental results show that the proposed algorithm can effectively identify the direction of motion blur with different fuzzy scales, and the estimation accuracy is higher than that of other similar algorithms, and the relative error rate is less than 0.02.
【作者单位】： 空军工程大学航空航天工程学院;
【基金】：陕西省自然科学基金(2014JM8332) 航空科学基金(20155896025)
【分类号】：TP391.41
，

本文编号：2306743