An Image Processing Techniques Used for Soil Quality Inspection and Classification
Mansur AS(a*), Herkules Abdullah(a), Hermawan Syahputra(a), Brahim Benaissa(b) and Fauziyah Harahap(a)

(a)Universitas Negeri Medan, Faculty of Mathematic and Natural Science, Jl. Willem Iskandar Psr V - Medan Estate, Medan, Indonesia.
*asmansur[at]unimed.ac.id
(b)Toyota Technological Institute, Department of Mechanical Systems Engineering, Design Engineering Lab, 2-12-1 Hisakata, Tempaku Ward, Nagoya, Aichi 468-8511, Japan.

Abstract

A soil inspection provides information on the essential fertility of the soil, and it is an important starting point for determining soil fertility. Therefore, soil quality determination is hazardous in agricultural systems before planting. Image processing techniques associated with the computer vision model are widely used today, having applications in many branches of agriculture, closely related to technologies used in precision farming. This study aims to provide an accurate model in image processing techniques for the inspection and classification of soil quality based on detecting external data. The outer texture was identified based on the visible and invisible system acquired using spectral technology (computer vision). We provided the Grey Level Co-occurrence Matrix (GLCM) method for analysis of the texture of images, and the classification process was performed using the Support Vector Machines (SVMs) method. This study showed that the model is a proper system for assessing soil quality. The experiment also indicates that the invisible channels have the potential in the classification model since the hidden texture features are not visible to the human eye.

Keywords: images processing- classification- GLCM- SVMs- soil inspection- spectral technology

Topic: Applied Sciences-Computers and Engineering