Transfer Learning based Recognition Algorithm for Common Tea Disease

Abstract

The technology of recognition for tea disease is help to increase the efficiency of disease control. On the basis of extensive analyses, a transfer learning based recognition algorithm for common tea disease (TLB_RA) was proposed to improve the identification accuracy of disease, such as anthracnose, tea blister blight and tea white scab. In order to improve the training accuracy, the inception v3 is exploited to built up the deep learning model under the condition of small sample set. Experiment results reveal that compared with the typical method, the recall level average and unit estimation of time is improved by the proposed algorithm.