Focusing segmentation method for light field refocusing image

Abstract

The invention discloses a focusing segmentation method for a light field refocusing image. The focusing segmentation method comprises the steps of transforming an input image into a neuron firing sequence space by using a PCNN computation model to generate a two-dimensional firing sequence matrix; in the firing sequence matrix, analyzing a connected region and a connectivity of each firing sequence number to obtain a connectivity attribute of each pixel; according to the pixel connectivity attributes and a pixel neighborhood firing sequence model, establishing a pixel classification criterion; for each pixel, marking a classification attribute according to the pixel classification criterion, and forming a classification mark matrix; and on the classification mark matrix, searching a region with the highest density of "clear marks" using a sliding window technique, and segmenting a corresponding region as a focusing region in an original refocusing image to be output. The method can obviously reduce the focusing segmentation time without losing the accuracy of the segmentation; and the method has simple configuration and good scalability, and is very suitable for the field of computationally intensive light field image processing.

Description

technical field [0001] The invention relates to a focusing segmentation method for light field refocusing images, belonging to the field of image focusing segmentation. Background technique [0002] As early as 1936, A. Gershun proposed the concept of light field to describe the radiation characteristics of light in three-dimensional space. However, the light field image acquisition technology lags far behind the development of the concept. It was not until 2005 that R.Ng et al. designed a hand-held light field camera, which greatly changed this situation. This creation promoted the rapid entry of light field cameras into the civilian market, and at the same time set off an upsurge in research on the application of light field images in academia. [0003] The light field camera can realize "photograph first, then focus", because light field imaging not only records the propagation position of light radiation, but also records the propagation direction. After the light fiel...

AI Technical Summary

Problems solved by technology

[0006] The focus segmentation of a single light field refocusing image is the basic step of multiple fusion focus segmentation, which can improve the effect of fusion segmentation of multiple images, but requires a time-saving lightweight calculation method, and a target segmentation method for low depth-of-field images does not meet this condition

Images