The Research And Implementation Of Tree Species Retrieval System Based On Tree Biological Characteristics

Computer science technology has been increasingly widely used in forest industry, intelligent retrieval of tree species is becoming more and more popular. Traditional identification and classification of tree species is mainly done manually, which is difficult due to the miscellaneous information of tree species. Users, especially those who work outdoor, are expecting a portable tree species retrieval tool that can help them minimize the input of the morphological characteristics and accurately identify a tree. So an intelligent mobile application that can improve the efficiency and accuracy of tree species identification, reduce the administrative burden of classified information is necessary and urgently needed.Based on the analysis of the existing retrieval systems and classification algorithms, and the comparison between different algorithms and their applicable data types, this article proposed a novel algorithm---Multi-valued Attribute and Single-sampled Data DecisionTree Algorithm(MSDT). Also, a new function is proposed to calculate the distance of two sets in comparison with commonly used functions. The experimental results show that the MSDT algorithm can efficiently address the entity identification problem for multi-valued and single-sampled dataset. The proposed new function applied to MSDT also shortens the steps and improves the performance.Moreover, this article designed and implemented a mobile tree species retrieval system based on Android smart phone platform. The system has a good test run:with MSDT algorithm as the model, this system greatly reduces the steps for retrieval and improves the chances to reach an unique result; good user interaction design helps to reduce user input errors, improve the accuracy of identification; supplemented by trees dictionaries and background management functions to achieve a dynamic query and update data in database.