Design And Application Of Greenhouse Environmental Regulation Rules Based On Decision Tree

With the development of the wireless sensor network and the rising of data mining technology,precision agriculture based on data driven has become a new trend.In the field of the greenhouse crop cultivation,the task of data mining is to mine decision rules from experts' cultivation experience for many years.Most traditional greenhouse control systems simply collect environment parameters,such as air temperature and humidity,their cultivation rules to control greenhouse environment are simply based on threshold,and they lack long-term assessment and effective feedback mechanism.This paper presents a novel spatio-temporal learning method to generate fine-grained cultivation knowledge of greenhouse crop from large environmental sensor data.Compared with human experience based control,the advantage of our method is that these fine-grained rules can provide precise climate control in time dimensions during the growth of plants.Spatial and temporal information are combined together and analyzed.The cultivation knowledge can guarantee similar quality of crop since it is especially important for the production.An unsupervised merging algorithm on time segments is proposed to learn optimal tags for classification,and then new data samples consisting of clustering labels are prepared for knowledge mining from the raw sensor data.Then,the climatic rules are built in the form of decision trees or decision cube.The experimental results show that almost 80%climatic conditions in the past successful cultivation can be copied to guide the future cultivation by our fine-grained decision rules.The advantage of our method is that farmers without expertise can easily implement temperature and humidity optimal control with the generated knowledge rules.Finally,based on the above theoretical research,an adaptive control method of flower greenhouse based on wireless sensor network is proposed and a hardware and software system combining sensing,decision-making and feedback is implemented.