A kinect-based matting method

Abstract

The invention relates to a Kinect-based image matting method. The method comprises steps: through OpenNI and OpenCV function libraries, the Kinect is used for photographing a target image, a color image and a deep image are obtained, and through a perspective conversion function in the OpenNI, the color image and the deep image are adjusted to be in the same perspective; with the help of the color image information, iterative image filtering is carried out on the deep image, and a foreground contour image is obtained; a ternary diagram is formed through a morphological corrosion expansion algorithm from the foreground contour image, and pixels in a given threshold range in the foreground contour image are defined as pixels in an unknown region; and a processing algorithm of removing slits and holes is carried out on the ternary diagram generated in the third step, and a final ternary diagram is generated. In comparison with the traditional image matting method, man-machine interaction in any form is saved, the Kinect is used for automatically generating the ternary diagram for image matting, and further in combination with the common sample point image matting algorithm, real-time image matting effects can be achieved.

Description

【Technical field】 [0001] The invention belongs to the field of image processing, in particular to a Kinect-based matting method. 【Background technique】 [0002] Keying technology is a very important image processing technology, which is used in the creation of film and television works, and also in medical treatment, agriculture and industry. However, under normal circumstances, the foreground part and the background part of the image often have coverage problems, that is to say, there are transparent pixels in the foreground part, so how to accurately separate the foreground part and the background part is the main content of the keying technology research. 【Content of invention】 [0003] In order to solve the above-mentioned problems in the prior art, based on the analysis of the traditional image filtering algorithm, the present invention proposes to use the depth image captured by the Kinect somatosensory device to key the color image captured by the Kinect. A new col...

AI Technical Summary

Problems solved by technology

However, under normal circumstances, the foreground part and the background part of the image often have coverage problems, that is to say, there are transparent pixels in the foreground part, so how to accurately separate the foreground part and the background part is the main content of the keying technology research.

Images