基于学习稀疏变换的工件降噪方法研究

摘要

Machine Vision is widely used for industrial manufacturing, such as detection, measuring, tracking, etc.Noise is the main interfering factor to the vision system.The method based on learning sparse transformation is proposed and the de-noising algorithm is modeled.Firstly, learning transformation is introduced, and solved by alternating between updating step and sparse coding step.Secondly, the algorithm is verified and the different index are evaluated.Lastly, the parameter is analyzed for the algorithm performance.The results show the algorithm can not only effectively filter the noise, but also proves to be superior in terms of running time.The method might be of great value of appliance for industrial manufacturing.%在工业制造中,噪声是干扰效果的重要因素.基于学习稀疏变换,提出了一种降噪方法,构建了降噪算法模型.首先,介绍了学习变换算法,建立了降噪算法模型,并分析了算法的交替求解法.其次,利用算法对实际工件图像降噪进行了验证,对不同指标进行了评价.最后,分析了参数对算法性能的影响.结果显示,提出的算法不仅能够有效滤除工件噪声,而且在运行时间方面有很大的优势,对工业制造的视觉应用有极大的应用价值.