Four-point transformation based image adaptive-correction method

Abstract

The invention discloses a four-point transformation based image adaptive-correction method, which is easy to implement and transplant, simple in operation, convenient to use and high in correction precision. The image correction method disclosed by the invention is implemented through acquiring double vision images, taking one of the double vision images as an original image, respectively acquiring four points on the original image and extracting an original image point; connecting the four points so as to form an arbitrary quadrilateral; through a mapping relation, obtaining corresponding four points, an original image point and an arbitrary quadrilateral on the other image and corresponding and traversing the image through double mice, thereby realizing the correction of the image, wherein in the process of building the arbitrary quadrilateral, a situation that no blind point or blind area occurs in areas outside the arbitrary quadrilateral is ensured by using an external expansion two-time method or an indentation two-point method, thereby realizing the adaptive correction, so that the image is corrected efficiently and accurately. The method disclosed by the invention can be applied to the field of image correction.

Description

technical field [0001] The invention relates to an image adaptive correction method based on four-point transformation. Background technique [0002] In the context of the rapid development of modern industry, many industries such as electric power, chemical industry, and metallurgy are using a large number of various cameras, including infrared and ordinary dual-view cameras. However, due to the different focal lengths of the two cameras, the position of the two lenses cannot overlap, and the parallelism of the axes and the complete synchronization of the installation angles cannot be guaranteed, the infrared images taken by them cannot be completely corresponding to the ordinary images. The difference between the two pictures makes it difficult to find the location of the equipment with extremely high temperature based on the infrared image. [0003] Existing image correction technologies (such as four-point straightening, homogeneous transformation and other arbitrary qu...

AI Technical Summary

Problems solved by technology

[0003] Existing image correction technologies (such as four-point straightening, homogeneous transformation and other arbitrary quadrilateral deformation technologies) almost all use complex algorithms such as eighth-order matrix and Gaussian elimination, which consume a lot of time and space complexity. And using multiple divisions, it is prone to unexpected situations such as the denominator being zero

In addition, many similar correction techniques only correspond arbitrary quadrilaterals to regular rectangles, and completely ignore the adjacent areas outside the quadrilaterals.

But in reality, there are very few cases where the image is inside a regular rectangle, and the scope of use is too narrow, and it often cannot meet the requirements in field practice.

Images