Detection and Recognition of Flower Image Based on SSD network in Video Stream

Abstract

At present, most flower images could only be recognized but not detected. They can only be used in the scenes with a single target instead of the scenes with two or more targets. Some application scenarios require the human-computer interaction mode with the current location information of flowers; moreover, due to the complexity of the environment and the similarity and difference between flowers, the traditional computer visual methods are inefficient and inaccurate. Therefore, this study introduced SSD deep learning technology into the field of flower detection and identification. The flower data set published by Oxford University was used as the research object, and it was used as the input of the neural network model for training and testing. The experimental results show that the average accuracy is 83.64% based on the evaluation standard of Pascal VOC2007, and 87.4% based on the evaluation standard of Pascal VOC2012. The processing time of an image on PC is 0.13s, which indicates that high-quality automatic detection and recognition can be performed, which can facilitate the retrieval of agricultural plant information database and help people to popularize related information of flowers.