The Study And Application Of Image Matching Based On Hausdorff Distance

Image matching has been an important topic in computer vision, object recognition, and image analysis. The performance of the matching method depends on the type of the features used, matching measure criterion, and searching strategy and so on. Most of the previous approaches to image matching can be categorized into the intensity-based and edge-based comparison. As a measure does not require point correspondences, hausdorff distance has been widely used for comparing point set or edge maps. The hausdorff distance measures the extent to which each point of a "model" set lies near some point of an "image" set and vice versa. It is reliable even when the image contains multiple objects, noise, and spurious feature. As a commonly used measure to determine the degree of resemblance between two objects that are superimposed on one another, it has been used to search images for instances of a model that has been translated or translated and scale. Genetic algorithms (GAs) are a heuristic search and optimization technique, initially developed by Holland. GAs is inspired by natural evolution, mimicking Darwin's principles of natural selection and survival of the fittest. In the part, they have been successfully applied to a wide range of real-world problems of significant complexity.In this paper, I propose a method for image matching based on hausdorff distance and genetic algorithms. I choose the edge of the image as the feature and the hausdorff distance as the similar metric. For the search strategy, I will choose the genetic algorithm. Some people has provided some methods for image matching based on the hausdorff distance and genetic algorithm, in which the model is only allowed to translate with respect to the image.In this paper, when using the genetic algorithm, I will first design a fitness function inspired by the method provided by Rueklidge for locating objects using the Hausdorff distance. Then, I will try some strategies combined with the common genetic algorithm mechanics to solve the image matching problem in the more general case of rigid motion, which includes translation, scale and rotation.