傅里叶和LBP描述子相结合的矿石颗粒种类识别  

Species identification of ore particles combined with Fourier and LBP descriptors

作　　者：罗小燕[1,2] 胡振 汤文聪 刘占 LUO Xiaoyan;HU Zhen;TANG Wencong;LIU Zhan(School of Mechanical and Electrical Engineering,Jiangxi University of Science and Technology,Ganzhou 341000,China;Jiangxi Province Engineering Research Center for Mechanical and Electrical of Mining and Metallurgy,Ganzhou 341000,China)

机构地区：[1]江西理工大学机电工程学院,江西赣州341000 [2]江西省矿冶机电工程技术研究中心,江西赣州341000

出　　处：《传感器与微系统》2023年第11期147-150,共4页Transducer and Microsystem Technologies

基　　金：江西省教育厅科学技术项目(GJJ200827)。

摘　　要：本文提出两种描述子相结合的方法进行矿石的识别,针对检测过程中矿石颗粒的运动与光照问题,局部二值模式(LBP)描述子具有良好的灰度不变特性,能够克服位移、光照不均等问题。傅里叶描述子只需少量的描述子,即可大致代表整个轮廓,从而能更好地区分不同的轮廓,进而达到识别物体的目的。分别提取两种算子,训练出两组异构数据,并利用判别典型相关分析算法得到矿石融合后的特征数据,然后利用K-means分类器进行矿石种类识别。在现场采集的矿石图像上进行试验,识别率高达88%以上。实验表明:提出的方法能有效提高矿石粒径测量的准确度。Aiming at the movement and illumination problems of ore particles in detection process,local binary pattern(LBP)descriptor has good gray-scale invariance characteristics,which can overcome the problems of displacement and uneven illumination.The Fourier descriptor only needs a small number of descriptors to roughly represent the entire contour,so that different contours can be better distinguished,and the purpose of identifying objects can be achieved.Two sets of heterogeneous data are trained by extracting two operators,and the characteristic data after fusion of ore is obtained by using the discriminant canonical correlation analysis algorithm,and then the K-means classifier is used for ore species recognition.The test is conducted on the ore images collected on site,and the recognition rate is as high as 88%.The final experiment shows that the proposed method can effectively improve the accuracy of ore particle size measurement.

关 键 词：图像处理 矿石颗粒 种类识别 边界角点序列 傅里叶描述子 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]