Automated soil clustering for crops cultivation

This paper focuses on the idea of cultivating a certain crop to a certain District-Sub District-Union of Bangladesh. The cluster and classification are done first on the basis of soil features and then comparing the soil features with the common features of the corps to be cultivated, is being searched for. Here k-mean Clustering, Fuzzy c-Mean clustering and SOM (Self-Organizing Map) based clustering algorithms are used. These algorithms are analyzed according to their results to find the better classification, by which one can find the most perfectly matched cultivable crops in a certain area. The big challenge of this paper is data collection, which was gathered from Sub-districts level books containing soil and crops features, both organic and physical property of Bangladeshi soil. These books were borrowed from Soil Research and Development Institute (SRDI), Khamarbari, Dhaka, Bangladesh. In this study, soil data of Louhogonj Sub-district of Munsigonj District in Bangladesh are analyzed.

I. Introduction

Bangladesh is a land of agriculture where a lot of people depend on farming. Regarding this if a better way can be introduced for crops cultivation, that will help this country. Soil clustering for better cultivation would be a great help in this sector. Clustering algorithms [1]–[5] have various applications such as soil clustering, food clustering, student group clustering, etc. Here it is used for soil clustering.