摘要
针对最小方向微分(MDD)算法在运动模糊方向识别中误差大和实时性差的缺点,提出一种改进的MDD算法。通过计算局部最大差值对模糊图像进行滤波,增强模糊方向上的纹理细节,利用双线性差值求解最小方向微分和,其对应方向即为模糊方向。根据方向微分和曲线的变化规律,给出模糊方向的迭代搜索模型,从而减少搜索次数。仿真结果表明,改进算法具有较高的识别精度和较快的执行速度。
Aiming at the shortcomings of big recognition error and many searching times by the Minimum Directional Derivative(MDD), a modified MDD algorithm is put forward. Local maximum difference of blurred image is calculated to enhance texture details in blurred direction. The blurred direction is obtained by the directional derivation summation curve. Based on the changing regularity of directional derivation summation curve, an iterative searching model which can reduce searching times is proposed to detect blurred direction. Simulation results show that this algorithm not only has high precision, but also executes fast.
出处
《计算机工程》
CAS
CSCD
2012年第6期175-177,共3页
Computer Engineering
基金
河南省重点科技攻关基金资助项目(102102210180)
河南省基础与前沿技术研究计划基金资助项目(092300410236)
关键词
模糊方向
纹理滤波
方向微分
双线性插值
局部最大差值
blurred direction
texture filtering
direction derivation
bilinear interpolation
local maximum difference