Abstract: Filtering denoising is not only the first step in seam image,but also one of the important contents to realize the vision seam tracking.To solve the problems existed in traditional median filtering,the affecting factors of median filtering were analyzed.The adaptive median filtering was introduced and the program was realized.Comparing with traditional median filtering,the adaptive median filtering could remove noise more effectively.Especially for the seam image of strong noise,processing performance of adaptive median filtering was far better than that of traditional median filtering.The results showed that noise of seam image could be effectively eliminated with edge perserving by adaptive median filtering algorithm,which laid a solid foundation for the subsequent processing.