| 窦燕*,王丽盼**,张启飞***.融合异质刺激过滤的金相图像等轴α相识别[J].高技术通讯(中文),2021,31(8):807~815 |
| 融合异质刺激过滤的金相图像等轴α相识别 |
| Metallographic equiaxed α-phase recognition fused with heterogeneous stimulation filtering |
| |
| DOI：10.3772/j.issn.1002-0470.2021.08.003 |
| 中文关键词: 金相图像识别； 异质刺激； 数学形态学； 标记分水岭算法 |
| 英文关键词: metallographic image recognition， heterogeneous stimulation， mathematical morphology， marking watershed algorithm |
| 基金项目: |
|
| 摘要点击次数: 1781 |
| 全文下载次数: 1146 |
| 中文摘要: |
| 针对钛合金金相图像存在噪声、晶界α相和片层组织与等轴α相颜色极为相似而导致“误识别”的问题，提出了一种融合异质刺激过滤的金相图像中等轴α相识别算法。该算法主要包含两大部分：（1）结合数学形态学和异质刺激理论设计实现了异质刺激模板，对晶界α相和片层组织进行了有效的过滤；（2）提出了一种结合距离变换和数学形态学的前景和背景标记方法，对等轴α相和其他组织进行了精确的标记，利用标记分水岭算法对金相等轴α相进行识别。实验结果表明，该算法过滤了金相图像中大部分的晶界α相和片层组织，并最大程度地降低了对等轴α相边缘区域的影响。与大津算法、最大熵算法、模糊C均值聚类算法对等轴α相的识别效果相比，本文算法提高了等轴α相的识别准确率。 |
| 英文摘要: |
| Aiming at the problem of ‘misidentification’ caused by the existence of noise in the metallographic image of titanium alloy and the fact that some of the grain boundary of α-phase and the lamellar tissue are very similar to the equiaxial α-phase in color, an novel equiaxed α-phase recognition algorithm fused with heterogeneous stimulation filtering is proposed. The algorithm mainly consists of two parts: (1) Combining with mathematical morphology and heterogeneous stimulus theory, the heterogeneous stimulus template is designed and realized to filter the grain boundary of α-phase and lamellar structure effectively; (2) A foreground and background marking method combined with distance transformation and mathematical morphology is proposed, and the equiaxial α-phase and other microstructure are accurately marked, and the metallographic equiaxial α-phase is identified by marking watershed algorithm. The experimental results show that the proposed algorithm can filter most of the grain boundary of α-phase and the lamellar structure and the influence of the edge region of the equiaxial α-phase is reduced to the greatest extent. Compared with Otsu algorithm, maximum entropy algorithm and fuzzy C-means clustering algorithm, the proposed algorithm in this paper improves the recognition accuracy of equiaxial α-phase. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
