基于图像清晰度评价的摄像头辅助调焦系统

Assistant camera focusing system based on image definition evaluation

提出一种新的图像清晰度(包括锐度和对比度)评价函数,对比传统的图像清晰度评价算法,新算法速度最快且准确,以及和功率谱算法一样优良的灵敏度,适合各种实时自动对焦系统.本文使用一系列不同程度的离焦图片演示了所有算法的性能。同时将新算法应用于一种实际的摄像头模块辅助调焦系统中,系统把采集的每帧图像实时显示在屏幕上,用新算法评价图像的清晰度,从而指导作业员对摄像头模块进行调焦。实践表明该摄像头辅助调焦系统在批量生产摄像头模块时对提高生产效率和稳定质量方面有很大作用。

A new evaluation function of image definition （include sharpness and contrast） is proposed in the paper. Compared with traditional methods, the algorithm is the fastest one and precise, and with good sensitivity similar to power-spectra algorithm. The algorithm is suitable to real-time auto focus system. The performances of all algorithms are demonstrated with a series of images with different degree of defocusing level. The algorithm is applied in an actual assistant focus system of camera module, which captures every frame of image and displays it on screen. By making use of the new algorithm to evaluate image definition, the operator can focus the camera module. Experience shows that this system is very helpful for improving efficiency and control quality during mass production of camera modules.