Damage detection in glass-fibre reinforced polymer(GFRP) structures, such as blades of wind turbines, is a challenging task to achieve using most of conventional methods used in Structural Health Monitoring (SHM). The primary cause of this issue is the relatively high internal damping of the material. Vision based methods however circumvent this issue. Among those methods hyperspectral imaging (HSI), a technique in which an image is recorded in a broad spectrum of electromagnetic radiation, has been proven to be a valuable tool for this purpose. Because of the high spectral resolution, hyperspectral images contain information about the chemical composition of the object being scanned. In this study, the chemical data contained in the hyperspectral images of GFRP samples is used as a basis for detection of presence of moisture-related damages. The aim of this study is to develop an algorithm allowing for detection of moisturerelated damage in GFRP structures. The algorithm utilizes the interaction of light with moisture through the phenomenon of absorption, cointegration analysis as a denoising and detrending tool, and machine learning methods for the purpose of classification. The results of proposed algorithm are evaluated and its applicability for the purpose of SHM is assessed.