Stitching multiple images into one panoramic photo
I first evaluated the 4 given images to compute the sequence. I calculate the sift features for all four images. The image sequencing is based on the number of matches. Additionally, ratio of matches to total features has to be greater than 20%. From the overlap matrix, I get the image that has maximum matches with the other images. I sort them according to the number of matches. I then get the index position for each image to finally get them in descending order of their number of matches. This gives me the sequence. Panorama After getting the sequence I pass the images in pairs to stitch them. This task involves computing the SSD (Here is set the ratio test value to 0.75) from the sift features, computing homograph, and warping and stitching. The resulting image is sent again to form a panorama. Also, since I increase the size of the matrix each time to increase the size of the background image, in the final step I also remove extra spaces in the panorama so that I get only the panorama cut out. Additionally, I also added padding to my images to make sure they don’t crop out of the frame.