An image-based moving object removal method

Abstract

The invention provides an image-based moving object removal method, which comprises the following steps: collecting N images at preset time intervals by a photographing device; preprocessing the image; extracting the eigenvalue of each image feature point, judging whether it is a foreground point or not according to the feature vector, and obtaining the foreground object frame by a connected domain algorithm; determining the position of all the foreground target frames of N images, selecting one of them as the fusion base map, and determine the needed patch map of each target frame; for each target frame of the fused base map, calculating the weight matrices of the fused base map and all patch maps performing the local illumination compensation on the patch map, and multiplying the weightmatrices of the base map and the patch map with each other respectively and then adding up to obtain the result map of the removed target. The method can be convenient for users to use the moving object removing function on the mobile device, input N images with moving objects, and finally obtain a composite image with only still images but no moving objects.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to an image-based moving target removal method. Background technique [0002] Existing moving objects can easily cause the partial loss of synthetic objects. Assuming that the moving object is a person, then in the final moving object removal map, it is very likely that a certain part of the human body is missing, such as feet, arms, For parts such as legs, this effect will seriously affect the user experience. At the same time, there are other synthesis methods, but they also have the problem of low resolution of the synthesized image. Contents of the invention [0003] Based on this, the purpose of the present invention is to provide an image-based method for removing moving objects. [0004] According to an image-based moving target removal method according to an embodiment of the present invention, the steps of the method include: [0005] Collect N images at prese...

AI Technical Summary

Problems solved by technology

[0002] Existing moving objects can easily cause the partial loss of synthetic objects. Assuming that the moving object is a person, then in the final moving object removal map, it is very likely that a certain part of the human body is missing, such as feet, arms, Legs and other parts, this effect will seriously affect the user experience

At the same time, there are other synthesis methods, but there is also the problem of low resolution of the synthesized image

Images