| 中文题名： | 基于Hough变换的粽叶图像自动定位与分割算法 |
| 姓名： | |
| 保密级别： | 公开 |
| 论文语种： | 中文 |
| 学科代码： | 070101 |
| 学科专业： | |
| 学生类型： | 学士 |
| 学位： | 理学学士 |
| 学位年度： | 2022 |
| 学校： | 北京师范大学 |
| 校区： | |
| 学院： | |
| 第一导师姓名： | |
| 第一导师单位： | |
| 提交日期： | 2022-05-23 |
| 答辩日期： | 2022-05-10 |
| 外文题名： | Automatic localization and segmentation algorithm for reed leaves image based on Hough transform |
| 中文关键词： | Hough变换直线检测 ; K-means聚类 ; FCM聚类 ; 分水岭分割 |
| 外文关键词： | Hough transform line detection ; K-means clustering ; FCM clustering ; Watershed segmentation |
| 中文摘要： |
数字图像处理是近来兴起的一项热门技术，其需求也越来越多。本文针对所给粽叶图像，为找出最上层粽叶并将其分割,提出了一种基于Hough变换的粽叶图像自动定位与分割算法：以Matlab为主要实现工具，对图像进行预处理后，运用Hough变换进行直线检测，找到粽叶的主叶脉，通过主叶脉定位最上层叶片位置，随后使用恰当的图像分割算法对其进行分割。该算法的难点在于叶片的定位与分离，在定位叶片的基础上，还需要将其与颜色信息高度相似的下层叶片分离。在分割方面，尝试使用了K-means均值聚类、模糊C均值（FCM）聚类、分水岭分割等算法。对图像处理后的结果显示，Hough直线检测算法在粽叶主叶脉的自动定位方面效果良好，并且准确率高、稳定性好。在三种分割算法中，分水岭分割达到的效果最好，保留了边缘缺口的信息，但仍需要进一步处理。
|
| 外文摘要： |
With the popularity and development of computer, digital image processing is becoming a popular technology that has emerged recently and its demand is increasing.In this paper, we propose an automatic localization and segmentation algorithm based on Hough transform for the uppermost leaf in the given image of reed leaves: Matlab is used as the main implementation tool, after the Preprocess, the Hough transform is used for line detection,to find the main leaf vein of reed leaves, the uppermost leaf position is located by the main leaf vein, after that the appropriate image segmentation algorithm will be used to segment it. The difficulty of this algorithm lies in the localization and separation of the leaves, and on the basis of locating the leaves, it is also necessary to separate them from the lower leaves with highly similar color information. For segmentation, algorithms such as K-means mean clustering, fuzzy C-mean (FCM) clustering, and Watershed segmentation were tried. The results of image processing showed that the Hough straight line detection algorithm worked well for automatic localization of the main leaf veins of dumpling leaves, and had high accuracy and stability. Among the three segmentation algorithms, the watershed segmentation achieves the best results and retains the information of edge notches, but further processing is still needed.
﹀
|
| 参考文献总数： | 10 |
| 作者简介： | 无 |
| 插图总数： | 16 |
| 插表总数： | 0 |
| 馆藏号： | 本070101/22152 |
| 开放日期： | 2023-05-23 |
谷歌
