Insect Recognition System Based On C/S Architecture

There are millions of insects found in nature,which are widely distributed all over the world.They are not only the largest animal population in the world,but also the largest number of individuals in each species.They are easy to breed and fast to reproduce.They are good experimental resources for scientific researchers.Research on insects covers the whole biological law,including genetics,ecology,physiology,biochemistry and morphology.Many natural science knowledges can be revealed by studying the characteristics of insects.For example,genetics developed by studying fruit flies.Many inventions in modern bionics are obtained by insects,for example,Fly-eye camera and vibration gyroscope invented by observing flies,artificial cold light invented by studying fireflies,helicopter invented by observing dragonflies and so on are numerous examples,which shows that the study of insects is of great significance to scientific development.However,there are many kinds of insects,which are not suitable for identification,and most of them are small.This makes it difficult for us to understand insects.With the high popularity of smart phones,not only brings many conveniences to people's lives,but also opens up a new way to learn knowledge.Users can quickly understand their professional knowledge by installing applications in related fields.Therefore,the design of an APP in entomology is of great significance to popularize the relevant knowledge.This paper designs and implements an insect recognition system based on C/S architecture,which consists of WEB server,Android mobile application and deep learning model.Using MySQL database to store user information and insect information,the basic functions of WEB server are realized by integrating three frameworks: Spring,Spring MVC and Mybatis.The system follows restful design style and uses HTTP protocol to realize data communication with client.WEB server is deployed on Ubuntu operating system.Tomcat is used as container.Client runs on Android operating system.Android studio is used and developed based on Android 8.0.Users can install the APP to learn about preset insects and identify insect species by uploading insect pictures.The realization of deep learning model relies on Keras high-level neural network API.TensorFlow deep learning framework is used as the back-end.A total of 19 small animal images of 16 common insects,such as dragonflies,longicorn,mantis and other similar insects,are selected.After repeated training,the recognition accuracy of the model reaches 98.5%.The development of this system involves many fields,such as website server,Android application,artificial intelligence,graphics,network programming and so on.At present,there are no similar software in the major mainstream Android application market.The development of this system provides a more convenient way for insect species identification.It can help non-professionals to understand entomology knowledge,familiarize themselves with insect species and distribution,enhance learning interest,enhance safety awareness and avoid accidental injury.At the same time,if applied to farmland,forest and other environments,it can help relevant staff to understand the habits of pests,grasp pest control methods,reduce the impact of pests and diseases,and have a positive impact on planting industry.This system is suitable for agricultural and forestry workers,insect enthusiasts,as well as primary and secondary school students.It can play a better role in campus education,knowledge popularization and other aspects.