Pest and plant disease identification in greenhouse using UAV images

Traditional methods on pest and plant disease prevention is visually checking the fields which is labor sensitive and time consuming. In addition, the relative knowledge to identify the pests and the diseases is required during the observation and checking. Therefore, the efficient protection of the crops and the improvement of the yields is still an issue. With the aids of various sensors and developments of artificial intelligent technologies, the crops can be monitored and the pests can be identified automatically. Based on these information, the farmers are able to make corresponding plans for pest prevention, which significantly reduces the labor efforts. Deep learning is a popular technique with various applications, such as object detection. It is possible using deep learning to detect the kinds of pest and calculate the amounts. There are many methods to detect the pest by using deep learning. However, most of the method need to obtain the images manually. The aims of this paper is to observe and detect the pest automatically. In order to get the images automatically, the way to obtain data is flying the Unmanned Aerial Vehicle (UAV) to collect the images from the greenhouse. After collection, deep learning is used to analysis and identify the amounts and the kinds of pests. This method could cost less time and manpower to look into the information of fields.