The performance of warp knitted fabrics as banner application is influenced by the structure of the fabric. Especially, the level of visibility of a banner fabric is dependent on the evenness of the fabric, which is also influenced by the characters of the component yarns. Until now, the status of fabric surface has been subjectively determined by human, which gives us inconsistent results. Thus, we try to study an objective method using image processing which characterizes the surface status of banner fabric. To do this, we prepare six different samples and extract parameters defining the surface status with image technology. Using the parameters, we analyze the relation between the surface status and yarn characters.