Terahertz (THz) technology is a competent non-destructive evaluation (NDE) technique, particularly for advanced materials such as Fibre Reinforced Polymer (FRP) composites due to its ability to penetrate most non-metallic and nonpolar substances. Typically, THz NDE studies are carried out using expensive and broadband pulsed THz systems limiting their widespread use in practical applications. In contrast, Continuous wave (CW) THz systems can potentially be a narrowband, cost-effective and scalable solution for NDE applications. However, conventional CW THz systems employ a coherent detection scheme which results in large acquisition time per pixel thus limiting their real-time applicability. In this paper, a CW THz system with incoherent detection scheme using Schottky receiver along with spatial adaptive sampling technique is employed to achieve rapid THz imaging of Glass Fibre Reinforced Polymer (GFRP) composite with artificial defects. Here, an initial coarse scan of 2 mm step size has been done, and gradient based thresholding criterion is used for identifying the regions of interest to progressively scan the sample with finer resolution down to a step size of 0.5 mm. Results demonstrate a total reduction in the image acquisition time by a factor of 50 compared to the coherent CW THz imaging. Further, the THz image acquired through adaptive sampling shows excellent correlation with that of the traditional uniformly sampled THz image with 0.5 mm step size. © 2019 SPIE.