All-in-One Focus image using Guided Filters

Abstract

Machine Vision System plays a vital role in manufacturing industries for the automatic inspection of quality of the finished product. The success of machine vision system for real time application is purely depends on the response of the hardware system in capturing the image of the object and the image processing algorithm which is used to extract the features of the object from the captured image. For the extraction of features, the image should contain the relevant information about all the objects in the scene and it should be noise free. Due to the finite focal length of the lens used in machine vision camera, it is possible to capture the clear image of objects in the scene which are in the particular range of distance from the lens. The other objects in the scene will be blurred. In this situation, image fusion is used to combine more than one source images with different focus to form the all-in-one focus image. In all-in-one focus image, all objects in the scene are in clear focus with greater visibility. One of the most widely used filters in image fusion to generate all-in- one focus image without noise while preserving edge information of object of interest is Standard Guided Filters. The time complexity of guided filters is O (N), where N is the image's pixel count.. Fast guided filters were proposed in the literature to reduce the time complexity of standard guided filters. This paper evaluates the performance of standard guided filters and fast guided Filters in fusing multi focused images to generate all-in-one focus image. Also this paper enhances the visibility of input images before image fusion process to improve the quality of all-in-focus image.