Abstract: Aiming at the problems in hand measurement staining area on cigarette surface,a measuring method based on image processing was proposed wherein a machine vision system was used as hardware platform.Firstly,a cylindrical surface model was used for image conversion to establish the mapping relationship between image coordinate system and expanding coordinate system.Then a tangent function was used for image enhancement to intensify the contrast between image and the background.Finally,the staining area was dissected by an improved regional growing algorithm to exclude the interferences of periodic texture and color distribution on cigarette.The results indicated that this method was of high measurement efficiency and stability with the precision of 0.03 mm²,and capable for determining stainings on cigarette.