Leaf Recognition Based on Color and Shape Using ADNet

To know the details of plants and leaves around us is of great importance medicinally and economically. Traditionally, plants are categorized by taxonomists through investigation of various parts of the plant. At the same time, most of the plants can be classified and identified based on the leaf shape and associated features. in this paper we USES Image processing techniques to extract leaf color, shape and GLCM features such as mean, standard deviation, homogeneity, correlation, aspect ratio, width ratio, moment ratio, apex angle, apex ratio, base angle, centroid deviation ratio and circularity. This paper proposes an adaptive visual tracking algorithm based on key frame selection and reinforcement learning (RL) Method. At the beginning, the probability value of the RL network output is analysed, and the predicted value of output is normalized. The proposed technic uses only a fine-tuned in key frames to obtain multiple fixed prediction models. The experiment is conducted on 75 video sequences of the Object Tracking Benchmark to verify the effectiveness of key frame selection strategy, and it is Compared with the original reinforcement learning based tracking algorithm.