首页|基于改进Canny算法和Hu矩的物表区域识别

基于改进Canny算法和Hu矩的物表区域识别

扫码查看
使用轮廓Hu矩匹配法进行物表检测的过程中,为解决轮廓匹配受噪声、比例变换等客观因素影响导致区域匹配识别率降低的问题,提出一种基于改进Canny算法和改进Hu矩的轮廓匹配方法进行物表区域识别。首先采用改进Canny算法降低噪声等对轮廓提取的影响,获取更优的轮廓信息,然后利用尺度归一法改进Hu矩进行物表区域的轮廓匹配,实现区域的匹配与识别。实验结果表明,该方法在噪声等因素干扰下可提取高精度的区域轮廓,且识别率至少达到98。3%,具有强鲁棒性和高识别率,满足物表检测对区域识别的高精度、高识别率要求。
Object Surface Region Recognition Based on Improved Canny Algorithm and Hu Moment
In the process of object surface detection using contour Hu moment matching method,in order to solve the problem that contour extraction and geometric moment are affected by objective factors such as noise and scale transformation,a contour matching method based on improved Canny algorithm and improved Hu moment is proposed for object surface area recognition.First-ly,the improved Canny algorithm is used to reduce the influence of noise on contour extraction and obtain better contour informa-tion.Then,the scale normalization method is used to improve Hu moment for contour matching of object surface region to realize re-gion matching and recognition.The experimental results show that this method can extract high-precision regional contour under the interference of noise and other factors,and the recognition rate is at least 98.3%.It has strong robustness and a high recognition rate and meets the requirements of high precision and high recognition rate for regional recognition in object surface detection.

object surface detectionarea identificationcontour extractionfeature matching

何坚强、翁嘉鑫、陆群、骆杨、蒋成晨

展开 >

盐城工学院电气工程学院 盐城 224000

江苏大学电气信息工程学院 镇江 212000

物表检测 区域识别 轮廓提取 特征匹配

国家自然科学基金青年科学基金项目

62003292

2024

计算机与数字工程
中国船舶重工集团公司第七0九研究所

计算机与数字工程

CSTPCD
影响因子:0.355
ISSN:1672-9722
年,卷(期):2024.52(8)