Focal plane positioning method, processing device, focal plane positioning system and storage medium

Abstract

In the focal plane positioning method, processing device, focal plane positioning system and storage medium of the present invention, the first depth image of the scene is captured by acquiring at least one depth camera located above the scene, and the second depth image corresponding to the height of the pedestrian's head is intercepted therefrom. image; detect each pedestrian head in the second depth image; locate the image coordinates of the reference point on the pedestrian head in the second depth image, and use the internal and external parameters of the depth camera to map to world coordinates; use the front camera The external parameter to map the world coordinate of the reference point to the coordinate under the camera coordinate system, obtain the depth data of the first focal plane where the reference point is located; obtain the distance from the reference point to the second focal plane where the face is located, and the offset distance is obtained The second focal plane where the face is located; the front camera is refocused on the second focal plane to obtain a corresponding face image; the present invention can realize precise focus extraction of faces with different depths in the same image.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a focal plane positioning method, a processing device, a focal plane positioning system and a storage medium. Background technique [0002] As an electromagnetic field distributed in space, light has various properties such as amplitude, phase, and wavelength, and helps humans perceive the brightness, position, and color of objects. However, traditional optical imaging intelligently captures the projected intensity of light radiation on a two-dimensional plane, while losing the three-dimensional shape data of the target. Light field data collection is to obtain the complete distribution of light radiation, and calculate the required image through data processing methods such as transformation and integration. Light field imaging refers to the process of collecting light field and processing the light field into an image. As a computational imaging technology, light fie...

AI Technical Summary

Problems solved by technology

For fixed scenes with a single application range, the focal plane can be manually set directly so that the light field system can be fixed on the same focal plane for light field rendering; for scenes that require foreground focus, image algorithms can be used to measure the scene foreground distance, so as to estimate the focal plane distance, the accuracy of this method largely depends on the effectiveness of the distance measurement method and the complexity of the scene environment; for the application of multi-focus plane requirements, the method of scanning focus can be used, that is Set different focus planes from far to near, know the light field focus, but this kind of focus method, first, has nothing to do with the scene content, so I don’t know which focus plane is accurate, and second, its effect depends on the scanning focal resolution

Images