Real-time wheat classification system for selective herbicides using broad wheat estimation in deep neural network

Abstract

Identifying seed manually in agriculture takes a long time for practical applications. Therefore, an automatic and reliable plant seeds identification is effectively, technically and economically importance in agricultural industry. In addition, the current trend on big data and data analysis had introduced scientist with many opportunities to use data mining techniques for better decision in various application. Recently, there are various number of applications that use computer-aided in improving the quality in controlling system. Classifying different types of wheat hold significant and important role in agriculture field. An improvement on such kind of system that makes distinctions based on shape color and texture of wheat plantation is crucial. The main objective of this paper is to develop a machine vision system which identifies wheat base on its location. For this purpose, a real time robotics system is developed in order to find plant in sorrowing area using pattern recognition and machine vision. For real-time and specific herbicide applications, the images are categorized in either expansive or precise categories via algorithm following the principal of morphological operation. Different experiments were conducted in order to gauge the efficiency of the proposed algorithm in terms of distinguishing between various types of wheats. Furthermore, the experiments also performed admirably amid varying field conditions. The simulation results show that the proposed algorithms exhibited 94% success rate in terms of categorizing wheat population which consists of 80 samples and out of them 40 are narrow and 40 broad.