A New Image Pre-Processing Algorithmic Based on HSV Color Space in the Process of Locating Container Characters

摘要

In this paper, a new Image pre-processing algorithmic based on HSV color space is proposed for locating container characters area to decrease the noise produced by non-container characters. Take advantage of hue value of the original container image in HSV color space to receive the hue value of the container body.. Process the original image in RGB color space to produce an initial binary image. Examine all the object pixels in that image, referring to hue of the surrounds of the pixels, to delete object pixels which not fall into container-body area. The experiments prove that the algorithm retaining container characters information while eliminating the most of the non-container area provides more convenience to completing image-analysis and locating container-characters area.