Image splicing method by means of blocking operation

Abstract

The invention discloses an image splicing method by means of a blocking operation. The method comprises the following steps of acquiring an original image which is photographed by multiple cameras, and a preset coordinate mapping table; based on the preset coordinate mapping table, determining a correspondence between a panoramic image and pixel coordinates of the original image, and dividing the original image into a plurality of small blocks with the same size by means of blocking operation; performing projection mapping on the pixel values of the pixels in each small block in the original block after blocking interpolation, and obtaining a final panoramic image according to a mapping result. According to the image splicing method, an operation is performed on the divided source image blocks and a correspondence is found through using each block as a unit. The pixel values of the source pixels are projected into target pixels according to the positions of the target pixels, thereby obtaining a panoramic picture which corresponds with a to-be-processed picture. The blocking operation based on a genetic algorithm reduces redundancy rate of the small blocks so that the processing speed of the panoramic image splicing method based on the blocking operation is greatly improved on an FPGA.

Description

technical field [0001] The invention belongs to the technical field of image communication, relates to the technical field of fisheye image splicing, and in particular relates to a method for real-time splicing of panoramic images using block operation. Background technique [0002] Fisheye image stitching is a technology that uses real-world images to form a panoramic space. There are two main schemes for current fisheye image stitching technology: [0003] The first solution is to use a conventional lens projection transformation and polynomial distortion correction technology to correct the lens distortion parameters with the correction template image, generate a set of intermediate corrected 2D images, and then use 2D image registration technology to correct the corrected The image elements are 2D registered, and finally the two registered images are fused. [0004] The second scheme is to combine the radial distortion correction parameters of the fisheye image and the ...

AI Technical Summary

Problems solved by technology

Although this method reduces the processing times of the computer, due to the bandwidth of the FPGA and the limitation of the cache, the processing efficiency of this pixel-based projection mapping on the hardware is very low

Images