FPGA Realization Method of Hamilton Adaptive Interpolation in Real-time Image Processing

Abstract

In order to realize the Hamilton self-adaptive interpolation algorithm on hardware logic, so as to perform high-quality color reconstruction on Bayer images, the invention provides an FPGA implementation method of Hamilton self-adaptive interpolation in real-time image processing. The present invention uses 6 rows of RAM to circularly buffer the original Bayer data, utilizes 2-way G component recovery parallel operation, 3-way R, B component parallel operation, pipeline processing, data synchronization and other methods, and realizes Hamilton self-adaptive interpolation algorithm on FPGA , can realize real-time processing of high-resolution (maximum support 4K*4K resolution) video streams, and obtain higher-quality images. There is a lot of margin.

Description

technical field [0001] The invention belongs to the technical field of real-time image processing, and relates to an FPGA realization method of Bayer image interpolation. Background technique [0002] A single CCD or CMOS color image sensor is formed by covering its surface with a color filter array (Color FilterArray). The image collected in this way is a Bayer image, and each pixel has only one component of R, G, and B. It needs to use a demosaic algorithm Carry out Bayer image interpolation calculation, restore each pixel into complete RGB three-component data, and finally obtain a high-quality image. [0003] Existing demosaicing algorithms can be roughly divided into two categories: [0004] The first type is the linear interpolation represented by the classic bilinear interpolation. This algorithm is easy to cause the loss and blurring of edge information. From the perspective of hardware, it can improve the throughput rate, but the restored image has poor subjective ...

AI Technical Summary

Problems solved by technology

[0004] The first type is the linear interpolation represented by the classic bilinear interpolation. This algorithm is easy to cause the loss and blurring of edge information. From the perspective of hardware, it can improve the throughput rate, but the restored image has poor subjective vision and is prone to jaggies. effect

[0005] The second type is the edge detection interpolation introduced by the direction operator represented by Hamilton and Kimmel. The introduction of edge detection restores the green component first, and then restores the red and blue. The quality of the reconstructed image is greatly improved, but there are Color moiré fringes, and because it is relatively complex and the restoration image process does not conform to the real-time processing of hardware logic, it is not easy to implement in hardware

Images