面向散斑纹理图像的相机聚焦评价算法

Camera Focus Evaluation Algorithm for Speckle Texture Image

摘要:

针对光学测量中操作员依照经验调节相机焦距而导致测量结果歧义这一问题进行研究，解决传统相机聚焦方法在散斑纹理图像上不适用的缺陷，基于灰度共生矩阵的熵、惯性矩和相关性三种特征量，提出一种面向散斑纹理图像的相机自动聚焦算法，辅助操作员调节相机。通过不同实验环境下采集的散斑纹理图像，验证所提算法单峰性和无偏性的特点，且具有一定抗噪能力。相比于传统评价算法，本文所提算法对散斑纹理图像的判断具有更强的针对性和更好的适用性，可为后续的精确测量提供保障。

Abstract:

Aiming at the ambiguity of measurement results caused by the operator adjusting the camera focal length according to experience in optical measurement, and solving the defect that the traditional camera focusing approaches could not applicable to speckle texture image, a camera automatic focusing algorithm for speckle texture image is proposed based on the entropy, moment of inertia and correlation of gray level co-occurrence matrix., which could assist the operator in adjusting the camera. The speckle texture images collected in different experimental environments verity that the proposed algorithm has the characteristics of unimodal and unbiased hardware, and has certain anti-noise ability. Compared with the traditional evaluation algorithm, the proposed algorithm has stronger pertinence and better applicability to the judgment of speckle texture image, and can provide guarantee for subsequent accurate measurement.