摘要
提出一种在基于图像的无损探伤检测系统中,图像二值化中的动态门限值的确定方法.通过对传统门限值确定法的分析,得出该种确定门限值方法在数学计算上的弊端,从而引出基于探测概率和虚警概率确定门限值的优点,并给出了基于概率与统计学上的推导过程.这种恒虚警概率条件下的动态门限值确定方法克服了传统方法的缺点,完全适应于计算机求解.因此,将该种理论与现在的处理技术和手段相结合,能够很好地满足在线检测系统的实时性要求.
In this paper, it clarified a method of the dynamic threshold determination of binary images in an imagebased non-destructive flaw detection system. According to the analysis of the traditional method to determine the threshold, it reached the threshold to determine types of mathematical methods in the calculation of the defects, thus leaded out advantages based on the probability of detection and false alarm probability threshold method to determine, and it was given that process based on probability and statistics was derived. It determined the dynamic threshold in specific conditions of the probability of false alarm overcomed the shortcomings of traditional methods, completely adapted to the computer to solve. The kind of theory and the combination of the current means of dealing with technology well to meet requirements of the on-line real-time detection system.
出处
《吉林师范大学学报(自然科学版)》
2009年第3期110-112,共3页
Journal of Jilin Normal University:Natural Science Edition
基金
省部共建教育部重点实验室项目(2006STS02)
关键词
图像处理
二值化
门限值
动态设定
image processing
binary
threshold value
dynamic setting