摘要
针对铜块合金样本感兴趣的区域(ROI)有效提取问题,提出了一种完整的单一背景下提取铜块合金样本ROI的方法,进而抽象ROI提取问题。将该问题分割为2个步骤:第一步为背景相关的种子点提取,第二步为背景无关的ROI确定,并提出了一种基于种子点的SBRE方法。复杂背景下的SBRE方法首先建立HSV颜色模型,然后通过不断更新步长收缩搜索矩形框最终使搜索矩形框完全位于铜截面内部,进而确定种子点,最终确定ROI。实际测试结果表明:该方法准确、快速。
Aiming at efficient extracting problem of copper alloy region of interest （l^OI） sample, a set of complete methods for copper alloy sample ROI extraction is proposed, extracts ROI problem. The problem can be divided into two steps：the first step is to extract seed point that is related to the background, and in the second step, the ROI which is unrelated to the background is determined, and further the SBRE method bssed on seed point is put forward. On the complex background, HSV color model is firstly established in SBRE method, and then the seed point is determined by locating the rectangle inside the copper cross-section through constantly updating step~ to shrink the research range and then seed point is determined, finally the ROI is determined. Actual test results nrove this method is accurate and fast_
出处
《传感器与微系统》
CSCD
北大核心
2013年第7期18-21,25,共5页
Transducer and Microsystem Technologies
基金
国家科技支撑计划资助项目(2011BAE23B05)
关键词
感兴趣的区域
种子点提取
SBRE方法
HSV颜色模型
收缩矩形框
region of interest（ROI）
seed point extraction
seed-based ROI extraction（SBRE）method
HSV color model
shrink rectangle frame
