Distance estimating device, distance estimating method, integrated circuit, and computer program

Abstract

A distance estimating device (1x) eliminates an uneven distance estimation accuracy that is a disadvantage when using a distance estimation method in which two images of an all-in-focus image and an arbitrarily focused image are used, and can obtain a high distance estimation accuracy for any subject distances. The distance estimating device (1x) comprises: a first generation unit (10w) for generating a first distance map showing depths specified from an all-in-focus image and a first depth focused image; a second generation unit (11w) for generating a second distance map showing depths specified from the all-in-focus image and a second depth focused image; and a distance map synthesizing unit (14) for synthesizing a distance map from the generated first and second distance maps.

Description

technical field [0001] The present invention relates to an apparatus and method for estimating distances to all subjects in images when capturing moving images or continuous still images. Background technique [0002] In general, distance estimation methods based on sensing can be roughly classified into the following two types. [0003] One is the active sensing type. For example, the active sensing type method includes: the time-of-flight method using laser light and ultrasonic waves; the confocal method commonly used in microscopes and the like; and the photometric stereo method using a plurality of light sources. [0004] The other is the passive sensing type. The passive sensing type methods include: a method of capturing a stereoscopic image and estimating the distance based on the difference in the positions of objects captured in each captured image; and a method of using different focal points of lenses while capturing one image simultaneously. [0005] The focus...

AI Technical Summary

Problems solved by technology

[0015] And, in addition to this, there is a problem that distance estimation accuracy can only be maintained with a relatively small blur

In other words, there is a problem that the accuracy of distance estimation decreases in areas where the change amount of blur is small and the blur is large.

Therefore, in general, the accuracy of distance estimation is lower than Coded Aperture and Coded Aperture Pairs

In addition, there is still the problem that the distance estimation accuracy can only be maintained with a relatively small blur

[0072] In the above-mentioned example, when the focus is near the near end, the distance estimation accuracy near the near end is high, but the distance estimation accuracy near the far end is low.

In actual use, the problem of deviation in the distance estimation accuracy is relatively large.

Images