IoT based Plant Monitoring and Identification using Low-Cost Image Sensors

This paper presents an intelligent crop monitoring system consisting of a controlled growing chamber, dimmable LED lights, imaging array, cloud data storage, and data processing components. The system is implemented using low-cost imaging sensors integrated with image recognition algorithms. A modified naive bayes classifier is developed for crop monitoring that estimates key color features of plants and utilizes them to identify plants as belonging to a specific plant type/class. The resulting data-base is continuously enhanced using the results of the analysis. Furthermore, the updated database is utilized to retrain the classifier periodically to account for the changing environmental and plant conditions. A series of experiments are presented that verify performance of the developed system.