Design And Implementation Of Control System In Plant Factory Based On Visual Patrol Robot

Since plant factories with artificial lighting(PFALs)were developed and commercially operated from 1970s,plant factornes were taken as the most promising growing technology and invested heavily in developed countries such as Japan and United States.PFALs are thought to be one of the most important part in agriculture because of many advantages such as high-efficient production,all-year-round production,free of pesticide and pests.Facing high risk of food hygiene and safety in China,PFALs are able to offer people fresh organic food.Therefore,lots of money have been invested to develop commercial PFALs.PFALs are facing many difficulties such as massive equipment costs,high operation costs and relatively small market share.The organic food produced in PFALs is much more expensive than food produced in open field,resulting in low profits of PFALs.Recent years,global PFALs research institues have managed to promote the automatic operation of PFALs through smart enviromental control system.They mainly focused on adjustment of LED spectrum and nutrient solution.With the costs reduction of LED lighting,better nutrient solution proportion scheme and smarter environmental control system,the operation costs of PFALs are reducing and become easier to make profits.However,it still takes employees a lot of time to check the plant's growing state and adjust the environment if necessary,which increases the employment costs.The visual patrol robot can identify the plant species and monitor its growing state,then adjust environmental factors.The robot can operate the PFAL smartlly and release employees from simple tasks,which would greatly promote PFALs' commerical development.In this thesis,an automatic operation system combined with a visual patrol robot is designed,whichi can classify diferent plants and monitor its growing state.This system can automatically adjustment environmental factors according to the plants' growing state.The robot analyzes the images of plants with photographic analysis algorithm and neural network model.The photographic analysis algorithm are proved to be effective and several neural network models are tested.Besides,researches on the mapping and navigation functions of the robot have also been done and integrated with the algorithms.Finally,an embedded system is designed as environmental control system,collecting sensor datas and adjust the environmental factors.