基于频率域运算的脉冲GMAW熔滴图像的线性处理
Linear image processing of pulse GMAW droplet photos based on frequency domain calculation
-
摘要: 通过微距高速摄影系统对脉冲GMAW过程进行拍摄,熔滴图像存在较多的焊接烟尘以及电弧光导致的环境噪声和衍射条纹,对进一步通过图像算法定量计算熔滴尺寸、缩颈变化等特征参数造成困难。因此在保证图像处理前后性质不变的条件下,采用频率域运算方式,设计了针对脉冲GMAW熔滴图像的衰减低通滤波器、高斯低通滤波器,并结合灰度映射和拉普拉斯叠加算法建立起了熔滴图像线性处理结构。结果表明,频率域的运算与线性处理结构的结合能够获得与噪声背景相分离的脉冲GMAW熔滴形态清晰图像,同时保证了图像的真实有效,为高精度定量分析脉冲GMAW过程提供了可能。Abstract: Based on micro focus high speed photography system, there are still lots of ambient noise and diffraction stripes caused by welding dust and arc light in the photos of the pulse GMAW (gas metal arc welding) droplet, which make the measurement of the droplet geometry, necking and other characteristic of droplets quantificationally more difficult. To ensure the property of the image before and after processing invariably, the linear image processing algorithm including the designed low pass attenuator and Gaussian low pass filter combined with gray level mapping and Laplacian addition was proposed based on frequency domain. The results of experiments show the combination of frequency domain calculation with linear image processing can obtain the clear photos of pulse GMAW droplet shape which separated from noisy background. The reality and availability of the photos is ensured. This method of also provides the possibility for pulse GMAW process high precision quantificationally analysis.