Depth estimation method based on focused image and image processing device

Abstract

The invention relates to a depth estimation method based on focused images and an image processing device, and the method comprises the steps of obtaining an image sequence, carrying out the superpixel segmentation of each focused image in the image sequence, and obtaining a plurality of superpixel regions; constructing a focusing evaluation curve of each superpixel region; and according to the focusing evaluation curve, calculating to obtain a depth estimation value of each pixel point, and forming a depth estimation image. According to the technical scheme, on one hand, superpixel segmentation is adopted, the segmentation boundary of the segmented superpixel region can coincide with the region, where the depth of the real object changes violently, of the real object, and therefore the problem that in the prior art, the region with the large actual depth change is excessively smoothed can be avoided; and on the other hand, the focusing evaluation curve constructed by the technical scheme has the advantages of wide steep region width and high definition ratio, and the accuracy of depth estimation and the anti-interference capability to noise can be ensured.

Description

Technical field [0001] The present application relates to the technical field of image processing, and particularly relates to an image processing method of the depth estimation apparatus based on a focused image. Background technique [0002] Three-dimensional depth estimation is crucial for computer vision applications, three-dimensional depth estimation as the field of computer vision is an important technology has been widely used by the two-dimensional image mapping, automated manufacturing, biomedical, autopilot and other fields. [0003] In order to achieve three-dimensional depth estimation, the prior art proposed focusing distance (depth from focus, referred to as DFF) method, the general process of the method can be summarized as follows: First, the focal length of the camera is acquired by changing parameters such as depth of focus objects Multiple focused image, calculates the focus function curve of the image pixels according to some function of focus, then the depth...

AI Technical Summary

Problems solved by technology

[0003] In order to realize 3D depth estimation, existing technicians have proposed a depth from focus (DFF) method. The general processing process of this method can be summarized as follows: firstly, by changing the camera focal length and other parameters to focus on the depth of the object to obtain Multiple focused images, calculate the focus function curve of the pixels in the image according to a certain focus function, and then obtain the depth estimation result of each pixel in the image; in some methods, it is further proposed to use linear weighting to improve the focus function To make up for the inaccurate focus estimation of the general method

[0004] However, the 3D depth estimation used in the prior art still has the following problems: (1) The inconsistency of the reference space of the image sequence cannot be avoided by changing the focal length of the camera for image acquisition, even if a telecentric lens is used, it will cause certain measurement errors ; (2) The imaging area where the depth of the surface of the object changes sharply is easily over-smoothed during the calculation of the focus function, resulting in inaccurate surface depth estimation; (3) The calculation of the focus function curve for each pixel on the image will cause the algorithm to Low efficiency, which is not conducive to the high-efficiency calculation requirements in practical applications; (4) When using the focusing function curve to calculate the depth estimation results, it is sensitive to noise interference, which seriously affects the results of 3D depth estimation

Images