An objective and reliable evaluation method for the efficiency of singeing process is developed by use of digital image processing. The overall fabric surface roughness before and after the singeing process has been estimated and evaluated comprehensively, specifically experimenting for darker shade samples that are prone to visual inspection errors. The energy, entropy, inertia, and homogeneity parameters of images of the fabric after and before singeing have been identified as the suitable descriptors of fabric surface roughness. A new integrating sphere imaging system that has been designed indigenously was used for image capture and analysis of unsinged and singed fabric with a digital DSLR camera. The imaging system ensures controlled, diffused and uniform illumination across the samples. The proposed evaluation method is found to be suitable and economic for analysis of singeing efficiency.