Specific Material Detection with Similar Colors using Feature Selection and Band Ratio in Hyperspectral Image

Abstract

Hyperspectral cameras acquire reflectance values at many different wavelength bands. Dimensions tend to increase because spectral information is stored in each pixel. Several attempts have been made to reduce dimensional problems such as the feature selection using Adaboost and dimension reduction using the Simulated Annealing technique. We propose a novel material detection method that consists of four steps: feature band selection, feature extraction, SVM (Support Vector Machine) learning, and target and specific region detection. It is a combination of the band ratio method and Simulated Annealing algorithm based on detection rate. The experimental results validate the effectiveness of the proposed feature selection and band ratio method.

Near‐infrared hyperspectral imaging for detection and quantification of azodicarbonamide in flour

“…Band ratio calculation is a type of spectral analysis and an effective image processing method. It not only reduces spectral difference between samples, but also increases contrast between differing samples . The method is widely applied for inspecting fruit quality, seed maturity and plant identification .…”

“…It not only reduces spectral difference between samples, but also increases contrast between differing samples. [26][27][28] The method is widely applied for inspecting fruit quality, [29][30][31] seed maturity 32 and plant identification. [33][34][35] In this study, band ratios were calculated to increase the contrast between flour and ADC, which was calculated as…”