Research and analysis on precise matching method for multi-feature of fuzzy digital image

In order to match the image features accurately, it is necessary to study the multi-feature matching method of fuzzy digital image. The current method cannot guarantee the efficiency and accuracy of image feature matching at the same time. In this paper, a multi-feature matching method of fuzzy digital image based on ACDSEE is proposed. This method combines the sparse matching method to match the image feature points, and uses the mean filtering, relative gradient, absolute gradient domain, image gradient domain enhancement model, K-means clustering algorithm, and pyramid principle to denoise, enhance, and segment the image, and extract features from the image before matching. Experiments show that the proposed method guarantees the accuracy and matching efficiency of image feature matching. © 2018, © 2018 Informa UK Limited, trading as Taylor & Francis Group.