@article{liu2019mature,
title = {A Mature-Tomato Detection Algorithm Using Machine Learning and Color Analysis},
author = {Liu, Guoxu and Mao, Shuyi and Kim, Jae Ho},
journal = {Sensors},
volume = {19},
nuber = {9},
pages = {2023},
year = {2019},
doi = {10.3390/s19092023}
}
A Mature-Tomato Detection Algorithm Using Machine Learning and Color Analysis
Introduction
An algorithm was proposed for automatic tomato detection in regular color images to reduce the influence of illumination and occlusion.
Framework
In this method, the Histograms of Oriented Gradients descriptor was used to train a Support Vector Machine classifier. A coarse-to-fine scanning method was developed to detect tomatoes, followed by a proposed False Color Removal method to remove the false-positive detections. Non-Maximum Suppression was used to merge the overlapped results.
Image Scanning Method
To deal with the tomato size variation problem, an image pyramid strategy was proposed for tomato scanning. Besides, an ROI was extracted using a Naive Bayes classifier based on the color feature of each pixel to reduce the search area and improve the efficiency.
False Color Removal
The False Color Removal method is proposed to eliminate the false positive detections by the SVM. The color feature of the sub-window is used to segment the sub-window. Then the white-pixel-ratio is applied to decide whether the window is a tomato or background. An optimization problem is solved for the color feature derivation.
BibTex
Related Work
Comment: Proposes an improved deep learning method based on YOLOv3 for tomato detection.