Shading judgment method and device based on depth layer

Abstract

The invention discloses a shading judgment method and a device based on a depth layer, and belongs to the field of computer vision. The method comprises the steps that a depth image of an original image is subjected to histogram counting; the number of pixels corresponding to all gray-scale values of the depth image is counted; envelope detection is performed according to a histogram; a depth counting envelope diagram is obtained, and subjected to neighborhood smoothing filtering; a minimum depth hop is computed according to the smooth depth counting envelope diagram; the minimum position translation quantity on a virtual view is computed accordingly, and serves as a cavity size threshold; and different cavity filling strategies are adopted for cavities with different sizes. The device comprises a depth histogram counting module, an envelope diagram computation module, a depth layer computation module, a cavity size computation module and a cavity size judgment module. The method and the device simplify the complexity of the existing rendering algorithm, accelerate rendering, and greatly improve the accuracy of cavity type judgment.

Description

technical field [0001] The present invention relates to the field of computer vision, in particular to a depth layer-based occlusion judgment method and device thereof. Background technique [0002] In recent years, with the continuous breakthrough of free viewpoint technology in hardware and software research, free viewpoint display has been favored more and more in the fields of medicine, exploration or entertainment. [0003] In order to make the effect of free view point display better, a variety of image rendering algorithms have emerged. Common rendering algorithms are roughly divided into two categories: [0004] The first type is the original image plus depth map rendering algorithm (2D plus Z), which is mainly based on the depth value corresponding to the scene point in the depth map, using the internal and external parameter matrix of the camera to calculate, first projecting and then retroreflecting to render a new virtual viewpoint view . [0005] The second ca...

AI Technical Summary

Problems solved by technology

[0009] (1) When rendering the image, neither of the above two types of methods classifies the holes; the holes are obviously divided into two categories, one is the small holes caused by the quantization error, usually a few pixels; the other is Depth jump is the large hole caused by occlusion, generally dozens of pixels

[0010] (2) When solving the hole problem, the first type of filling technology can well deal with the small holes caused by the quantization error, because it cannot fundamentally solve the hole problem, especially when filling the larger holes caused by occlusion, it will be obvious Distortion, the filling effect is not ideal; the second type of smoothing technology deforms the object, although the number of large holes is reduced, but the visual quality of the obtained virtual viewpoint is significantly reduced

Images