改进的图像纹理检索方法在矿石识别中的应用.pdfVIP

11 11 V o.l 11, N o. 11 2006 11 Journa l o f Im age and G raphics N ov. , 2006 1) 2) 2) 1) 陈 玲 沈红标 李咸伟 刘其真 1) 2) ( , 200433) ( , 20 1900) , , , , , , : TP391. 41 : A : 1006-896 1( 2006) 11- 1700-04 Application of Image Texture Retrieval inM ineralD iscrim ination 1) 2) 2) 1) CHEN L ing , SH EN H ong-biao , L I X ian-w e i , L IU Q -i zhen 1) (D ep artmen t of Comp u ter S cien ce, F udan Un ivers ity, Shanghai 200433) 2) ( In stitu te of Env ironm en t and R esource, B aosteel C o , L td , Shangha i 201900) Abstract T he p ap er dep ict s m inera l discr mi ina tion through mi age texture retrieva l. The w ho le procedure goes like th is: f irstly d iv id ing a large mi age into sm a ll w indow mi ages, w ho se fea ture va lu es are then ex tracted according to coo ccurrence m atr ices and gray leve l run leng th m a tr ices, w ith the lmi it o f practica l com puta tion capab ility, on ly one sing le fea ture vec to r is app lied in the c lassif ica tion. A s w e can see in the p ap er, a satisfy ing consequence can be reached w hen an appropr ia te threshold value can be prov ided. K eywords texture recognition, feature vec to r, smi ilar ity , , 1