Fruit Yield Estimation Using Shallow Neural Network

Additional Funding Sources

The project described was supported by a Specialty Crop Block Grant of the Idaho State Department of Agriculture.

Abstract

One of the tools for precision agriculture is yield monitoring. In this paper, a yield monitoring system using machine vision is developed to estimate fruit yield early in the season. Predicting yield early in the season helps farmers in the marketing of their product and the production logistics. The machine vision system uses a color camera to acquire images of the trees during the blossom period. An image segmentation algorithm was developed to recognize and count the blossoms on the tree. The segmentation algorithm used a shallow neural network that used the color information and position as input. There was a high correlation between the blossom count and the number of fruits on the tree which shows the potential of this method.

Comments

T20

This document is currently not available here.

Share

COinS
 

Fruit Yield Estimation Using Shallow Neural Network

One of the tools for precision agriculture is yield monitoring. In this paper, a yield monitoring system using machine vision is developed to estimate fruit yield early in the season. Predicting yield early in the season helps farmers in the marketing of their product and the production logistics. The machine vision system uses a color camera to acquire images of the trees during the blossom period. An image segmentation algorithm was developed to recognize and count the blossoms on the tree. The segmentation algorithm used a shallow neural network that used the color information and position as input. There was a high correlation between the blossom count and the number of fruits on the tree which shows the potential of this method.