SIFT算子与改进算法检测木板条纹的对比研究  

Timber Strip Detection Based on SIFT and Improved Histogram Algorithm

作　　者：张发军 宋钰青 邓安禄 虞成俊 彭文刚 ZHANG Fajun;SONG Yuqing;DENG Anlu;YU Chengjun;PENG Wengang(College of Mechanical&Power Engineering,China Three Gorges Univ.,Yichang 443002,China;Hubei Key Laboratory of Hydroelectric Machinery Design&Maintenance,Yichang 443002,China)

机构地区：[1]三峡大学机械与动力学院,湖北宜昌443002 [2]水电机械设备设计与维护湖北省重点实验室,湖北宜昌443002

出　　处：《三峡大学学报（自然科学版）》2020年第1期98-102,共5页Journal of China Three Gorges University：Natural Sciences

基　　金：国家自然科学基金项目（51275273）;水电机械设备设计与维护湖北省重点实验室开放基金项目(2017KJX02)

摘　　要：为了解决在木板加工分类中的分级问题,运用图像处理技术,提出一种检测木板条纹间隔大小的改进直方图算法对木板进行检测分类,并将该算法得到的特征点与SIFT(Scale-invariant feature transform)算法进行对比,试验结果明显看出改进直方图算法比SIFT算法更有效,更能准确识别条纹木材的条纹,准确率最高可达90%以上,且算法更为简单.In order to solve the problem of classification of timber,an improved histogram algorithm for detecting the strips of timber is designed by image processing technology.Comparing the results between the algorithm with the SIFT(Scale-invariant feature transform)algorithm and that proposed in this paper,it can be concluded that the improved histogram algorithm is more efficient and more accurate.The accuracy rate is up to 90%,and the algorithm is simpler.

关 键 词：木板条纹 SIFT算法 改进直方图算法 计算机视觉 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]