摘要
植物根系原位CT图像的分割是植物根系三维重建和定量分析的基础。在综述了有关领域常见图像分割算法的基础上,分析了植物根系原位CT序列图像分割研究中存在的问题,提出了原位根系CT序列图像分割算法研究的新思路。指出原位根系CT序列图像分割算法的研究应针对CT序列图像的特点、结合根系空间构型的先验知识,综合利用数学形态学、模糊理论以及神经网络等理论方法,重点在提高算法的精度和鲁棒性。
The precise segmentation of roots CT images is the basis to implement the 3D reconstruction and quantitative analysis of plant root system in situ. The main methods of image segmentation used in the relevant fields are summarized. The characteristics and ap- plicability of these methods are discussed. The problem existing in the research on the CT image segmentation for plant roots in situ are analyzed. Some new ideas are put forwards for the further research. The characteristic of CT slice images and the transcendental knowledge of roots＇ geometrical architecture should be considered fully in the new algorithms of root CT image segmentation. And the new relevant theory and methods including mathematical morphology approaches, fuzzy mathematics and neural network-based approaches should be used synthetically in the new algorithms. Its emphases are to improve the precision and robustness of algorithms.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第17期4252-4256,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(60375005)
