The rapid development of digitalization has challenged the inheritance of intangible cultural heritage, and in recent years, non-heritage elements have made considerable development in the field of clothing design. In this paper, we develop the non-heritage characteristic element extraction system, design the element extraction program, construct the image extraction model of non-heritage virtual clothing design, and extract the characteristic elements such as pattern, color, outline and other characteristic elements in non-heritage culture through K-means algorithm and morphological algorithm. The non-heritage elements are integrated into the virtual clothing design through 3D clothing modeling. After evaluating the image processing performance of this paper’s model, further color analysis of non-heritage elements in clothing design is carried out by taking shadow play elements in clothing design as an example. The model of this paper, with SSIM and PSNR values of 0.887 and 24.68 on the training set, and 0.815 and 23.95 on the test set, far exceeds other models. Among the shadow elements of the virtual costume, the main color of the old Dan element is yellow, accounting for 67.75%. The main colors of the young Dan are yellow and green, and the hues are concentrated in H ∈ [10°, 113°]. The dominant colors of the clown role are yellow and lime green. The main colors of both raw and clean roles are red and blue-green, with a color range of H ∈ [1°, 213°]. The colors of all five roles are dominated by medium-high saturation, with the highest saturation in Xiaodan. The color brightness of the five roles is between the ranges of [3, 100].