What splicing skills are there to eliminate large screen splicing?

Mosai is an image fusion technology, which generally refers to the fusion of the same type of images. He put overlapping information blocks together as a part of image cohesion, a complete picture, and deleted redundant information. Image mosaic technology is widely used. For example, virtual reality is widely used in panoramic display, remote sensing image processing and other fields. The evaluation criteria of image mosaic should not only have good visual effect, but also keep the spectral characteristics of images as much as possible. Generally speaking, mosaic images are smoother and have better effects. Of course, "seamless" is not absolute, but "seamless" in the range of human resolution.

In general, when splicing large screens, seams will inevitably appear on the splicing boundary. This is because slightly different gray levels in the two images to be stitched will lead to obvious stitching seams. Moreover, in the actual imaging process, this subtle difference is inevitable. Therefore, the difficulty of image mosaic technology lies in finding the position relationship between images and smoothing the two images together to get a global image. The basic idea of this paper is that by breaking the previous method of finding the stitching line, as long as an optimal stitching point is found, a straight line becomes unreasonable stitching, but a closed value is taken within the value range of each point to find the point and the line is used as the stitching line for stitching.

Method of eliminating flat-fell seam

1, median filtering method to eliminate joints median filtering method is to carry out median filtering in the area near joints. The difference between the pixel value and the surrounding pixels is close to the value of the surrounding gray value, thus avoiding the intensity discontinuity of the light intensity. Near the narrow-band median filter processing connector. This method is fast, but the quality is average. Smoothing results will reduce the resolution of the image, make the details of the image unresolved, and cause the image to be blurred.

2. Wavelet transform is a common method to eliminate joints. It makes full use of the characteristics of multi-resolution wavelet transform and solves the problem of image mosaic well. The principle is: because wavelet transform has the characteristics of band-pass filtering and the bandwidth actually occupied by wavelet transform components at different scales is large? The higher the frequency of the components, the smaller the bandwidth of each wavelet component. The first of the two images is decomposed into different frequency components of wavelet according to wavelet decomposition method. As long as there is a sufficiently fine decomposition mosaic, the wavelet component of the bandwidth can be small enough. Then, under different scales, two images are selected according to different sizes, and then the wavelet components of the images are put together first, and then the whole image is restored by the restoration program. Images can well balance sharpness and smoothness. However, wavelet transform also has its shortcomings, such as the complexity of wavelet transform algorithm and the need for splicing in the first wavelet transform domain. A large number of floating-point operations and boundary processing problems are involved in the calculation process, which makes the actually produced large-capacity image with large memory overhead of clock counting machine take a long time to process and the stitching speed is slow.

3. Weighted smoothing method is used to eliminate joints. In practice, there are many methods to use weighted smoothing for overlapping areas. The idea of this method is to average the gray values of the corresponding points in the two images by weight to get the gray values of LPixels in the overlapping areas of the images, that is, Pixel-kxlpixel+(l-k) xrpixel, where: k is a gradient factor and meets the conditions.