The Enhanced RBF-BFOA Network for Remote Sensing Image Classification

One of the primary objectives of remote sensing data processing is to extract maximum relevant information for the user. Optimization plays a crucial role in computer science, especially in learning problems such as neural network training. In this paper, we introduce a new bio-inspired algorithm to optimize the classical Radial Basis Function neural network: the Bacterial Foraging Optimization Algorithm (BFOA). The proposed approach is first presented and described, followed by an enhanced algorithm through modification of the chemotactic step in BFOA. Subsequently, we assess and compare the performance of both approaches through various remote sensing image classifications. After different tests, it is found that the proposed approach is very robust and efficient.