Interpolation and Sharpening for Image Upsampling

This paper presents a series of studies and discussions on upsampling of low-resolution images to obtain high-resolution images from interpolation methods. Firstly, the traditional nearest neighbor interpolation, bilinear interpolation and bicubic interpolation methods are reproduced and optimized respectively. Then different sharpening algorithms are discussed to solve the problem of blurred edges. By compensating the contours of the image, thus making the image sharp. The fourth-order and eighth-order Laplace operators are used respectively. Meanwhile, this paper refers to the idea of RCAS in FSR technology in order to find the Laplace operator that makes the best effect. Subsequently, the lanczos4 function is used to further solve the problem of distortions caused by the traditional interpolation method. The lanczos interpolation and linear interpolation methods are also applied to the non-edge and edge of the image, respectively, to achieve the best clarity. Finally, by combining two different sharpening algorithms with Lancozs algorithm, the images are upsampled with less distortion as possible. The software algorithm is also implemented on FPGA for verification.