According to the color yarns in the fabric, the fabrics can be divided into three categories: solid color fabrics, single-system-melange color fabrics, and double-system-melange color fabrics. The density of solid fabrics can be inspected with gray-projection method or Fourier analysis method. But the methods cannot be applied to yarn-dyed fabrics directly. A method for detecting the density of single-system-melange color fabrics will be discussed in this article. By analyzing the pattern and color characters of single-system-melange color fabrics, fuzzy C-means algorithm is proposed to classify the colors in the fabric image based on CIELAB color space first. With the color segmentation results, the fabric can be divided into different blocks. The yarns can be located in different blocks with different average gray-levels, and then the number of yarns can be counted in each block. The linear density of threads can be obtained by counting the yarns in a unit length finally. The experiment proved that the algorithm proposed in this study can inspect the density of single-system-melange color fabric successfully. (c) 2012 Wiley Periodicals, Inc. Col Res Appl, 38, 456-462, 2013