高温轴类大锻件的图像分割算法研究

采用非接触CCD图像测量技术对高温大锻件进行测量时,需要从采集的图像中分割出大锻件目标图像。对高温物体成像特性进行了研究,并对高温锻件的辐射出射度与成像灰度的关系进行了讨论;采用线扫描方法,获得了850～1000℃高温锻件倒棱滚圆过程中图像的RGB基色分布规律,发现了背景和锻件目标区域红色灰度均值和蓝色灰度均值的相关性。在此基础上,提出了一种基于动态相对阈值和动态绝对阈值的图像分割算法。实验表明,此方法能对图像中的锻件目标实现精确的分割。

Research on image segmentation algorithms for high temperature shaft heavy forging

The image of heavy forging object should be segmented when measuring the high temperature heavy forgings based on non-contacted CCD image measurement technology.The characteristics of high temperature object image as well as the relationship between thermal radiation of high temperature heavy forgings and image grays were analyzed.The distribution of RGB primary colors in images of high temperature heavy forging from 800 to 1000 ℃ during chamfering-rolling process as well as the relativity between the averages of red gray and the average of blue gray both in background and forging object zones were obtained through the line scanning method.A kind of image segmentation algorithm based on dynamic relative threshold and dynamic absolute threshold was put forward.Experimental result shows that the heavy forging object can be segmented precisely by the proposed segmentation algorithm.