Level Sets Method and Corner Detector Fused Approach to Image Classification

This paper presents a framework for detecting objects in images. The motive of this research can be found in onsite requirements. We focus on the practical needs on distinguishing salt called purity from impurities, which are sand, soil, and other substance in the heaped salt on a conveyer belt. Our corner detector is given by an auto-correlation function about images. In this work, basing on image energy, we formulate the auto-correlation function on image energy to construct a piloting set which includes possible elements to classify objects and the impurity object in order to navigate the front propagation of the level sets. We finally use level set method to detect the topologic changes of curves and to catch the objects/impurity. The available has been proved by applying this approach into an industrial image shown in this paper.