Method of Prediction on Vegetables Maturity Based on Kalman Filtering Fusion and Improved Neural Network

The morphology of vegetable will change significantly in the process of growth. This paper uses image segmentation technique and edge detection technique to extract canopy leaf area, plant height, root length, lateral area of root and other external morphological characteristics. Then train the characteristic parameters of aerial and underground parts based on neural network in order to realize the prediction on the vegetable maturity. Furthermore, we newly propose a method to fuse the two parts of information and predict maturity based on the Kalman filtering, which has been proposed to improve the accuracy. The results show that neural network can be used to realize the prediction of maturity, and there is an improvement on the accuracy which is up to 97.5% after using Kalman filters to fuse the information on maturity of upper and latter part of the plant.