基于改进JSEG技术的茶叶图像嫩芽分割与识别研究  

Research on Tea Leaf of Image Segmentation and Recognition Using Improved JSEG Algorithm

作　　者：方坤礼[1] 廖建平[1] 刘晓辉[2] 

机构地区：[1]衢州职业技术学院机电工程学院,衢州324000 [2]浙江大学机械工程学院,杭州310027

出　　处：《食品工业》2017年第4期134-138,共5页The Food Industry

基　　金：衢州市科技计划项目(2014Y016);国家自然科学基金项目(50975255);浙江省教育厅科研项目(Y201432278)

摘　　要：提出一种新的彩色纹理图像的分割技术(JBSEG),该技术基于边缘检测方向算子的幅值来调节JSEG,克服JSEG和HSEG的局限性,同时兼顾局部区域相似性和边界非连续性。试验表明,与JSEG和HSEG技术的分割结果相比,该技术能更好地避免过分割现象的发生,同时结合超绿特征2R-G-B能有效分离茶叶中的嫩芽。A novel segmentation algorithm JBSEG for color images is proposed on the base of JSEG. The method improves JSEG based on amplitude of directional operators for edge detection in images, and overcomes the limitations of JSEG and HSEG methods to take account of both homogeneity and discontinuity in regions. Compared with image segmentation results of JSEG, the technique can avoid the occurrence of over segmentation on large amounts of images, the tender leaf quickly identifies using super green feature algorithm. Finally, the proposed method has better performance of segmentation and recognition of tender leaf.

关 键 词：图像分割 JSEG技术 JBSEG技术 茶叶 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]