Smart farm prototype for plant disease detection, diagnosis treatment using IoT device in a greenhouse

摘要

One common problems faced by farmers is to identify the diseases in plants. Just relying on naked eyes to detect diseases is very difficult. We propose the development of a cost-effective and affordable smart farming prototype that can detect plant diseases in advance and notify farmers so that remedial actions could be taken. Also, a web-based system has been developed to allow the farmers to monitor the status of their plants in the greenhouse. The development of such a smart farming plant disease detection system will allow farmers to have the ability to detect diseases at an early stage of the plants which will minimize yield loss from full-blown disease outbreak in their greenhouse. One important feature of this system is its ability to categorise the Septoria plant disease for which categorizing the stages of the disease outbreak on the leaf is very important. Furthermore, we have demonstrated that machine learning classification algorithm can be used in raspberry pi without losing performance. Based on our tests we also find that random forest classifier with colour histogram is the best combination to be run by this prototype for efficient plant disease detection.