Colour is a subjective perception, and in an industrial environment using colour, objectivity is of great significance. In pursuit of improvising the measurement approaches inheriting the scientific progressions, this research emphasis on the development of an automated system with image processing and machine learning techniques for non-contact colour assessment of both prints and textiles under user-defined daylight conditions in the light booth. The system consists of a light booth with tunable LED daylight luminaire to set the day-lighting conditions of D50, D65 and D75 with adjustable illuminance as per colour assessment standards of ISO/ASTM. The feature vectors of the sample images are extracted using colour histograms through histogram quantization. Colour classification is performed using K-Nearest Neighbor (KNN) algorithm trained with 140 shades of Macbeth colour checker chart SG. The proposed system is compared with visual and instrumental measurement methods for experimental validation. The results demonstrate an accuracy rate of 86% in colour classification of prints. Correlating to the lightness of textile samples an accuracy rate of 86% (very dark colour), 83% (medium light colour), 100% (very light colour) found