摘要
直线是图像分析过程中非常重要的描述符号。在工业控制中,图像处理通常采用最小二乘法对直线进行拟合,但在对估计精度要求较高时,传统最小二乘法往往不能满足要求。将离散小波变换和传统最小二乘法相结合,建立了一种基于小波预处理的最小二乘估计的新方法,获得了比传统最小二乘法效果更好的估计结果。试验证明了该方法的有效性和高精度性。
Straight line is a very important descriptor in image analysis process. In industrial control, the least square method is commonly used for fitting of straight line in image processing. When higher accuracy of estimation is requested, traditional least square method cannot satisfy the requirement. Thus, the discrete wavelet transform is combined with traditional least square method, to establish the new type of solution based on wavelet pre-processing for the least square estimation, to acquire better estimation result than that of traditional least square method. The experiments verify the effectiveness and high precision of this method.
出处
《自动化仪表》
CAS
北大核心
2012年第2期19-21,24,共4页
Process Automation Instrumentation
基金
甘肃省教育厅科研基金资助项目(编号:00330715-01)
关键词
最小二乘法
小波变换
阈值处理
小波重构
图像处理
Least square method Wavelet transform Threshold processing Wavelet reconstruction Image processing
