Flower Species Recognition Based on the Convolutional Neural Network

Abstract

At present, in botany, agronomy, and species research, if the identification and classification of flowers is only done manually, it may require a lot of manpower and the recognition accuracy rate is low. Moreover, traditional computer vision and artificial intelligence are widely used. The search query in the database also faces some problems of high recognition cost, low recognition rate and low efficiency. In response to these problems, this article uses five common flower image data sets based on the deep learning field and image information processing problems. Based on the convolutional neural network framework, the flower processing is divided into the following four processes: image information import, preprocessing, Feature extraction and classification of image information. Through the training and verification of five kinds of flower data set graphics, the recognition accuracy rate of this model reaches 75%, and the recognition accuracy rate is improved compared with traditional recognition methods.