Feature Detection and Matching with Lineous Adjustment and Adaptive Thresholding

In this work, feature matching is performed using the FAST feature descriptor. The identification of image features and matching technology are key elements of the view in computer vision.

Even then, the problem remains between fast response and stable matching in real-time. We suggest a method for extracting image features and matching using image processing techniques to solve this issue. To solve this problem, we adopt rotation, contrast adjustment, and blurring for image pairs processed by medium filtering. 

Here, the FAST approach is applied for feature selection to improve the performance. The proposed method is performed on a bike dataset which is available on online sources (Google). Particularly, this paper focuses on the illumination change, image blur, and image rotation aspects.

Keywords: Image feature identification, Feature matching, Feature descriptors, Feature extraction, FAST feature.