Essential processing methods of hyperspectral images of agricultural and food products

Authors: JIA, BEIBEI - China Agricultural University; WANG, WEI - China Agricultural University; Ni, Xinzhi; Lawrence, Kurt; Zhuang, Hong; Yoon, Seung-Chul; GAO, ZHIXIAN - Tianjin Institute Of Agricultural Resources And Environmental Sciences

Submitted to: Chemometrics and Intelligent Laboratory Systems
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/9/2020
Publication Date: 1/11/2020
Citation: Jia, B., Wang, W., Ni, X., Lawrence, K.C., Zhuang, H., Yoon, S.C., Gao, Z. 2020. Essential processing methods of hyperspectral images of agricultural and food products. Chemometrics and Intelligent Laboratory Systems. 198:Article 103936. https://doi.org/10.1016/j.chemolab.2020.103936.
DOI: https://doi.org/10.1016/j.chemolab.2020.103936

Interpretive Summary: Hyperspectral imaging (HSI), also known as " imaging spectroscopy" or "imaging spectrometry", combines spectroscopic and imaging techniques in one imaging modality to get both spectral and spatial information simultaneously. A hyperspectral image is composed by thousands or even millions of pixels. This technology has been widely utilized in quality and safety assessment of different agricultural and food products, such as fruits, vegetables, beef, lamb, pork, poultry carcass, fish, cereals, and dairy products. Hyperspectral images integrate spatial and spectral details together. However, the collected data also included background noise. This paper focuses on the currently available pre-processing and post-processing methods that have been successfully used in the analysis of hyperspectral images of agricultural and food products. Three types of pre-processing or pre-treatment of hyperspectral images and two post-processing of hyperspectral images described in this paper can be performed to improve the classification result of images or to generate chemical images /distribution maps to show spatial component concentration distributions of non-homogeneous samples.

Technical Abstract: Hyperspectral images integrate spatial and spectral details together. They can provide valuable information about both external physical and internal chemical characteristics of agricultural and food products rapidly and non-destructively. Despite rapid improvements in instruments and acquisition techniques, the collected high-quality hyperspectral images still contain much useless information (like uneven illumination, background, specular reflection, and bad pixels) that need to be removed. Thus, hyperspectral image pre-processing is essential to reduce negative influences on the subsequent steps of detection, classification, and/or prediction analysis. This manuscript focuses on the currently available pre-processing and post-processing methods that have been successfully used in the analysis of hyperspectral images of agricultural and food products. For pre-processing or pre-treatment of hyperspectral images, methodologies are further divided into three types: image-based, spectrum-based, and spatial-spectral interactive methods. Moreover, post-processing of hyperspectral images can be carried out to enhance the classification result of images or to generate chemical images /distribution maps to show spatial component concentration distributions of non-homogeneous samples.