Date of Publication :12th July 2017

Abstract: — Real-time performance is a critical demand to most of the applications, which require the detection and matching of the visual features in real time images. This process establishes the correspondence between two images taken at different time. The proposed hybrid SIFT features are invariant to change in scale, rotation and intensity. This algorithm has been used in navigation for detection of ships and planes, used in traffic controls and detection of objects in many applications.The hardware architecture to detect stable keypoint module and descriptor generator module has been designed. The input image is driven from a digital camera for real time implementation and the pixel values are recorded for the key point detection. The key points are obtained using FAST corner detector and the images are matched based on the key point obtained. The edge response rejection module is used to eliminate edge features of the image which makes the key-points unstable. After the detection of stable key points, the orientation and magnitude are calculated for each pixel to perform image matching. In the proposed hybrid SIFT, FAST key points and SIFT descriptor recognizes the image in an effective and efficient manner. The hybrid SIFT minimize the time due to less key points and the rotation delay is reduced above 60 0 . The proposed hybrid SIFT has been verified using MATLAB R2013B for various images. The hardware architecture has been simulated using ISim p2.8xd and tested for various images.

Reference :