Automatic focusing method based on calibration information

Abstract

The invention relates to an automatic focusing method based on calibration information. The automatic focusing method comprises two steps of imaging system calibration and rapid search focusing. Imaging system calibration is realized by performing primary defocusing-focusing-defocusing whole-process image acquisition through Gaussian fitting of a definition evaluation value and position information; the quick search focusing is to perform image acquisition analysis in a far-focus area and a near-focus area, perform Gaussian fitting on image definition and position information, and take a fitting position as a focusing position. According to the method, the invariant characteristics in hardware are analyzed and quantified through imaging system calibration, a follow-up search focusing strategy is strictly planned globally, and the efficiency and stability are greatly improved. The automatic focusing method provided by the invention is good in unimodality, high in timeliness and high inadaptability, meanwhile, the focusing frequency is greatly reduced, and sigma/8 high-precision focusing can be realized only by collecting 5-6 points in a near-focus area.

Description

Technical field: [0001] The invention belongs to the technical field of automatic focusing of cameras, and in particular relates to an automatic focusing method for quickly searching a current field of view based on the advance calibration of an imaging system. Background technique: [0002] The camera's auto-focus technology is widely used in industrial testing equipment, measurement photography, medical imaging equipment and other fields. The methods of automatic focusing can be mainly divided into two categories: active measurement focusing methods and focusing methods based on image definition. [0003] The active measurement focusing method refers to adding a measurement device outside the imaging system, or embedding the newly added measurement device into the imaging system structurally. The measuring device used in the active measurement focusing method measures the distance of the detected object. When the distance measurement value exceeds the depth of field allow...

AI Technical Summary

Problems solved by technology

[0039] 1) The selection of the initial large step size is generally tested or roughly estimated, and it is not easy to grasp. If the initial step size is too small, the search efficiency will be reduced, and if the initial step size is too large, the near-focus area will be skipped, resulting in failure to focus

[0040] 2) A single detection of the falling edge is to reverse the direction, which is susceptible to noise interference, causing the algorithm to fall into a local extremum or turn back at a position deviated from the maximum value, and the focusing effect is not ideal

The existing improved hill-climbing method uses multiple (3 or more) falling edges in a row to judge the reentry, which can reduce the noise interference, but will greatly increase the calculation amount of the algorithm

Images