Optimal Approach for Enhancement of Large and Small Scale Near-Infrared and Infrared Imaging

Abstract

In a broad area of industry such as remote sensing and medical diagnosing, imaging enhancement technology takes a leading role, where energy distribution of the light source depends not only on image coordinate but also on wavelength. Both infrared (IR) and near-infrared (NIR) imaging techniques have a variety of applications in these fields. For instance, satellite images are taken via IR or NIR spectrometer and laser Doppler medical scanning is collaborated with NIR spectrometer. Matrix functions of any image correspond to brightness or energy at each image pixel. The actual decision making must rely on detailed investigation of images being obtained. Therefore, image processing should be taken into account so as to enhance the results from real world. Segmentation is an image analysis approach to clarify feature ambiguity and information noise, which divides an image into separate parts that correlate with the objects or areas of the particular object involved. This procedure can be conducted by clustering, which is a process of partitioning a set of pattern vectors into subsets. Being a simple unsupervised learning algorithm, K-means clustering algorithm has the potential to both simplify the computation and accelerate the convergence. In most cases optimization is closely related to clustering, which gives rise to the best way of problem solving. In this article, optimal approach is proposed to be implemented along with image segmentation. This methodology is to enhance both large scale and small scale IR and NIR image processing.