摘要
水下焊接图像经阈值处理后求取各个光斑的平均高度,利用贝叶斯最小概率模式识别对V形光带的左右边进行分类。获得左边一类后记录最左边点的水平坐标,并求取这些点的平均高度;同理,获得右边一类后记录最右边的一个点的水平坐标,并求取这些点的平均高度。以这2个平均高度作为最终的基准高度,结合V形焊缝的深度,即可最大限度地消除V形焊缝上下的干扰光斑,以最左边和最右边的水平坐标为界可以消除焊缝左右边的干扰。
At first,making threshold to an underwater V-groove weld to get every facula's average height,then the left side and the right side of the weld were classified by the minimum probability of Bayes.Using this way could calculate the average height of these dots in the left of the weld and note the horizontal coordinate of the leftmost dot in the left of the weld,the right of the V-groove weld could be determined and the horizontal coordinate of the rightmost dot could be noted,and the average height of these dots in the right of the weld could be obtained as the same way.The average of the two average heights was calculated as a benchmark,combining the depth of the V-groove weld with the benchmark,the faculae could be removed in a weld image that lie in up and down the word of 'V' laser light-strip as possible as it could,and could also remove the faculae in it that lie in the left and right side of light-strip with the leftmost and rightmost horizontal coordinates.
出处
《焊接技术》
北大核心
2009年第9期39-41,共3页
Welding Technology
基金
国家863项目资助(2007AA04Z242)
江西省科技攻关项目资助(2007BG09100)
关键词
贝叶斯
最小概率
模式识别
光斑
V形焊缝
Bayes,the minimum probability,mode detection,faculae,V-groove weld