Image correction method and system, electronic equipment and storable medium

Abstract

The invention discloses an image correction method and system. The method comprises the following steps: obtaining a depth map and an RGB image of a shooting target; setting at least one feature region on the depth map and / or the RGB map; acquiring a depth value of the feature region and a corresponding position of the feature region on the depth map or the RGB map, and calculating to obtain a target transformation matrix; and correcting the RGB image according to the target transformation matrix and parameters of shooting equipment. The RGB image is corrected based on the depth map, so that an accurate RGB image can be obtained.

Description

technical field [0001] The invention relates to a data processing method, in particular to an image correction method and its system, electronic equipment and a storage medium. Background technique [0002] Deep learning is an algorithm based on representation learning of data in machine learning. Observations (such as an image) can be represented in a variety of ways, such as a vector of intensity values ​​for each pixel, or more abstractly as a series of edges, regions of a specific shape, etc. However, in recent years, deep learning has attracted more and more attention, and has been applied in many imaging fields, including face recognition and human body posture recognition. However, most of the image processing technologies based on deep learning are based on three primary color (RGB) image processing technologies. Therefore, the disadvantages of the existing technologies are high training costs, poor real-time performance, and high requirements for the complexity of ...

AI Technical Summary

Problems solved by technology

But there is no guarantee that the object being photographed is also in a vertical state

For human body modeling, if the person leans forward or backward when taking pictures, or the camera is not vertical, it will cause the person to be out of parallel with the image plane and introduce perspective errors.

Images