改进Canny算子在工业影像清晰度评价中的应用研究

Study of an Improved Canny Algorithm Using in Industrial Image Definition Evaluation

【目的】为了实现不同离焦和正焦图片边缘检测中高低阈值的自适应确定。【方法】提出一种基于改进Canny算子的自动对焦清晰度评价函数。该函数利用传统Canny边缘检测算子的运算迅速、抗噪性强和精度高的特点,将模糊熵应用于Canny算子高低阈值的自适应确定,并用图像边缘点的模糊熵和作为图像清晰度评价值。【结果】在工业影像测量自动对焦系统中,算法的有效性得到了验证。【结论】试验结果表明,清晰度评价函数在满足单峰性和无偏性的前提下,具有较高的实时性,以及对环境较强的适应性,能够满足工业测量系统中对自动对焦的需求。

[Purposes]Auto-focusing is one of key technologies in the industrial image measurement system.[Methods]It presents an auto-focusing definition evaluation function based on improved Canny algorithm.The function has characteristics of quick operation,stronger noise resistance and high precision of Canny algorithm,the fuzzy entropy was used to determine the high threshold and the low threshold adaptively,and sum of the fuzzy entropy about edge points was used as the definition evaluation value.[Findings]Effectiveness of the algorithm was verified in the auto-focusing system of industrial image measurement.[Conclusions]Experimental results indicated that,under the premise of satisfy unimodality and unbias,the definition evaluation function has high real time capability,and high adaptation.This algorithm can satisfy the requests of the auto-focusing in the industrial measurement system.