In order to solve the problem of glisten and noise, a six-step pretreatment method was proposed in this paper. To solve the problem of glisten, area of glisten should be confirmed by threshold selecting. Then glisten area was filled by pixel of background. And then dial scale was resumed by character interpolation. Guassian filter and median filter were used in wipping off Guass noise and salt-pepper noise. Laplace sharping was used in improving edges and details. Experimental results show that six-step pretreatment method can decrease glisten and noise in image of pointer instrument. Six -step pretreatment proposed a new idea for pretreatment of image of pointer instrument, and it has some referenced value.%为了解决指针式仪表图像中的反光和噪声问题,提出了一种六步预处理方法.对于反光问题的解决,先通过阈值筛选确定反光区域的位置,再用背景像素加以填充,最后用特征插值恢复表盘刻度；对于噪声问题,通过高斯滤波和中值滤波去除高斯噪声和椒盐噪声,再利用拉普拉斯锐化环节改善滤波造成的模糊.实验结果表明,六步预处理方法有效地削弱了指针式仪表图像中的反光现象和噪声,为指针式仪表图像的预处理工作提供了一种新的思路,具有一定的参考价值.
展开▼
