一种基于改进ZS细化算法的指针仪表检测

发布时间：2018-04-01 16:34

  本文选题：指针仪表　切入点：颜色特征提取　出处：《计算机工程》2017年12期

【摘要】：针对ZS细化算法存在二像素畸变、细化不彻底等问题,提出一种改进的细化算法。对二值图像采用ZS细化算法进行初步细化,将初步细化后的结果进行全细化识别,依据识别结果对未完全细化的图像采用4连通细化模板实现彻底细化。同时对细化过程可能出现的断点问题,采用构造保留模版的方法加以解决,进而实现目标图像的有效细化。以油量表为例,结合改进细化算法实现仪表读数识别。仿真结果表明,改进细化算法在保持原有ZS细化算法快速性的基础上可有效地解决二像素畸变、像素冗余的问题,并实现准确的指针读数识别。
[Abstract]:Aiming at the problems of two-pixel distortion and incomplete thinning in ZS thinning algorithm, an improved thinning algorithm is proposed. According to the recognition result, the 4 connected thinning template is used to realize the complete thinning of the incomplete thinning image. At the same time, the breakpoint problem that may appear in the thinning process is solved by the method of constructing the reserved template. Then the effective thinning of the target image is realized. Taking the oil meter as an example, the instrument reading recognition is realized by combining with the improved thinning algorithm. The simulation results show that, The improved thinning algorithm can effectively solve the problems of two-pixel distortion and pixel redundancy on the basis of keeping the speed of the original ZS thinning algorithm and realize accurate pointer reading recognition.
【作者单位】： 江南大学物联网工程学院;
【基金】：江苏省六大人才高峰基金(2012-WLW-006)
【分类号】：TP391.41;U463.7
，

本文编号：1696351