This paper focuses on the use of imaged-based machine learning techniques for identifing grain. In particular we compare several texture descriptors based on Local Binary Patterns(LBP),and we report new experiments using a set of novel texture descriptors based on the combination of the Elongated Quinary Pattern (EQP), the Elongated Ternary Pattern (ELTP) and the Elongated Binary Patterns(ELBP).These three variants of the standard LBP are obtained by considering different shapes for the neighborhood calculation and different encodings for the evaluation of the local gray-scale difference. The resulting extracted features are then used for training a machine-learning classifier(support vector machine). Our results show that a local approach based on the EQP feature extractor, which can express both local and holistic features of the grain image, produces a reliable system for identifing grain.

Date Published: 4 February 2011
PDF: 8 pages
Proc. SPIE 7752, PIAGENG 2010: Photonics and Imaging for Agricultural Engineering, 77520O (4 February 2011); doi: 10.1117/12.886165

Published in SPIE Proceedings Vol. 7752:
PIAGENG 2010: Photonics and Imaging for Agricultural Engineering
Honghua Tan, Editor(s)