Research Of Tree Species Classification And Recognition Algorithm Based On Leaf Image Characteristics

As floorboard of woody plants,there are many kinds and shapes of tree,and tree is widely distributed in the wilds and cities.The traditional method to identify trees is manual query mainly by detailed characteristics recorded in?Flora of China?and?Tree Records of China?,which not only wastes time and energy,but also requires researchers to master enough forestry and botany foundation.Faced with so many kinds of trees,it is very difficult for experienced forestry researchers to identify the tree species quickly and accurately,which blocks the development process of dendrology and trees classification.Therefore it's expected by forestry experts and the public to develop a digital tree species identification tool with high accuracy,large portability,easy use and high profession.This paper regards tree leaf as the main research object and designs an image preprocessing algorithm of eliminating leaf shadow.This paper self-defines two leaf outline characteristics of leaf angle and edge angle mean on the basis of 25 kinds of characteristics such as leaf texture,invariant moments and traditional shapes,regards the definition and inference of similar polygon as theoretical basis and proposes a kind of classification method for tree species identification which is to construct distance matrix and corner matrix based on leaf outline.The algorithm has been implemented on Android Studio platform by JNI hybrid programming.In Flavia dataset,the accuracy of the algorithm is 99.61%among 1907 normal samples of 32 categories.In Leafsnap dataset,the accuracy of the algorithm is 98.26%of top 5 among 23147 Lab samples of 185 categories.