基于小波变换的自适应阀值植物根系图像边缘检测
Roots image edge detection based on adaptive thresholds and wavelet transforms
-
摘要: 植物根系的生长状况可以反映该地区的气候及土壤特性,现有的根系研究方法如挖掘法、整段标本法和剖面法,都有破坏样本和工作量大等缺点.为了实时跟踪植物根系生长状况,介绍了一种基于多尺度小波变换和自适应阀值的图像处理方法,对采用内窥方法获得的根系图像进行边缘检测,并将边缘提取后的图像进行融合.此方法可以在不损伤植物根系的前提下自动对根系的生长情况进行提取分析,实现实时采集及精确测量根系的物理参数.Abstract: The root's growing condition can reflect the climatic characteristics and soil status. Existing methods such as excavation, monolith and bisection will destroy the sample and have too much work to do. To track the root's growing addition in real time, this paper introduced a method based on adaptive thresholds and multiscale wavelet transforms, by which we can detect the edge of a root's image that obtained by endoscopy and then mosaic the images extracted from the edge. The method analyses the growing condition of the root automatically without any destroying. Through this system, we can acquire real-time collection and accurately measure the root's physical parameters.