A pre-processing method for obtaining differential image

Abstract

The invention provides a preprocessing technology for obtaining differential images, in particular, for two input digital images with specific deformations, the horizontal and vertical stretching and compression deformations, rotation deformations and translation deformations of the images are gradually restored to realize image Overlap, and finally get the method of two pictures that can be differentiated. The occurrence of specific deformation means that only one or more of horizontal and vertical tension-compression deformation, rotation deformation and translation deformation occurs in the image, and the horizontal and vertical tension-compression deformation occurs last. It is characterized in that two images are input, regardless of whether there is relative deformation between the two images or what relative deformation is, the second image is sequentially restored for the three deformations of horizontal, vertical compression, rotation and translation, so that The common parts of the two images are completely overlapped, and then image difference is performed to output an image that highlights the different parts of the common parts of the two photos. Using this method, the input photos can be automatically compared, restored and differentiated to obtain a clear differential image.

Description

Technical field [0001] The present invention relates to a preprocessing technology for obtaining differential images, in particular to two input digital images that undergo specific deformations, and gradually perform image restoration processing of horizontal and vertical tension and compression deformation, rotation deformation and translation deformation, and realize the image Coincidence, and finally get two different pictures. The occurrence of specific deformation means that the image has only one or more of horizontal and vertical tension and compression deformation, rotation deformation and translation deformation, and the horizontal and vertical tension and compression deformation occurs at the end. With this method, images can be automatically compared and restored quickly, and the images preprocessed by this method can be differentiated to obtain clear differential images. When it is necessary to take multiple photographs of the same object, often due to the inevitable ...

AI Technical Summary

Problems solved by technology

[0003] 1. Image difference can only be performed on images that are completely overlapped and have no deformation

[0004] 2. It is only limited to the recovery process of translation and deformation of the image, and then image difference

[0007] 1. The images that can be differentially processed have certain limitations, and the differential images must be completely coincident with the reference image or only have translation deformation;

[0008] 2. Using the method of numerical calculation, although it avoids the restoration process of image deformation, it also brings new problems such as complex mathematical calculation and slow processing speed

At the same time, the image may also have various forms of deformation such as distortion and radial distortion, and this method is not suitable for dealing with these forms of deformation

Images