Research Of Disease Pest And Weed Identification System For Mobile Terminal

Disease, pest and weed are natural enemies of crops. They make a significant impact on agricultural production. Disease pest and weed are various and complex. There are significant difficulties in identification and prevention. The current identification of disease pest and weed is mainly by comparing professional profiles artificially. This identification method requires high professionalism, but has low efficiency and poor durability. This topic do some research on disease pest and weed ’identification system based on Android. The performance of the self-development system is a collection of a variety of intelligent information technology in one smart identification system, which can achieve real-time automatic identification of disease pest and weed. The main research contents are as follows:1. The thesis does some research on the disease pest and weed recognition system, and discusses the current status of expert systems and image recognition technology. Then it points out the shortcomings of existing similar studies. Next, it describes the major platform of smartphone, and the principle of an Android developing system, JSON format and the basic flow of the image recognition.2. Then, this thesis describes in detail the demand analysis and the outline design. The technology solution presented here includes the design of hardware and software platforms, the overall system architecture design and communications mechanism design. The thesis also does system requirements analysis and functional design.3. This thesis makes a rather detailed analysis of disease pest and weed recognition algorithms. It analyses and contracts the principle and application of the algorithm. It then discusses an algorithm of disease pest and weed classified identification. The algorithm extracts 3 kinds of features, as known as color, texture and shape features, from the disease pest and weed images.4. Based on test result, this system can recognize disease, pest and weed form image with high accuracy in short time. This system can be applied to practical application.