Related Points and Geometric Constraints Based Straight Line Matching Method

To combine related points and line can increase the line's features description ability. However, straight lines exist error matching situations. In order to improve line matching accuracy, a straight line matching method based on related points and geometric constraints is proposed. Firstly, related points are chosen and used to construct affine invariant to describe line feature. And lines are matched coarsely based on the similarity of affine invariant. Then, affine transformation model is calculated and straight lines are matched finely by the constraints of line angle and distance. The experimental results show that the method is suitable for image with affine transformation images, can obtain more matched lines pairs with high accuracy. It can obtain good line matching results for images with affine transformation and images with complex or simple texture.