稻穗产量性状对水稻生长发育、遗传育种及功能基因解析研究具有重要意义，而传统的有损获取方式存在脱粒误差大、实粒瘪粒区分难的缺点。因此本项目针对稻穗产量性状无损解析展开研究，旨在不脱粒，不分离实粒、瘪粒的情况下，快速、准确获取稻穗产量性状如有效穗、穗粒数、结实率、千粒重、一次枝梗数、二次枝梗数、及着粒密度。本项目提出采用可见光、X-ray双模式成像方式，同步获取稻穗的可见光彩色图像和X-ray透射图像，利用可见光图像产量特征解析技术如椭圆检测、分水岭分割、并行细化处理算法获取单颗穗粒区域、稻穗分支骨架，利用X-ray图像解析技术获取单颗米粒区域、穗颈节区域图像特征，再将两幅图像特征进行配准融合，基于穗粒轮廓及内部米粒特征解决实粒、瘪粒的区分问题，并建立稻穗产量性状的数学表征，实现稻穗产量性状的无损提取。本研究工作将为水稻产量性状获取提供一种创新的途径，并可能促进水稻研究的进一步发展。

Rice panicle yield-related traits have great significance to the study of growing development, genetic breeding, and functional gene research. However, the traditional method need to thresh the panicle and separate the filled spikelet from the unfilled spikelet, which are inaccurate and time-consuming. Therefore this research proposes a nondestructive measuring method without threshing and spikelet separation, in order to obtain the rice panicle yield traits accurately and fast. The imaging modalities based on visible light and X-ray, are adopted in this study, which are used to acquire the color image and X-ray transmission image of rice panicle synchronously. Then the color image is analyzed to extract the area of each spikelet and panicle branching skeleton, and the X-ray image is also processed to extract the area of each grain kernel and panicle neck feature. After that, the image registration is applied to build the feature fusion image, by which the filled spikelet and unfilled spikelet are distinguished. Finally, the mathematical characterization of rice panicle yield-related traits are set up, to extract the panicle yield-related traits. In conclusion, the development of this study will provide a new method for the panicle yield traits acquisition, which will probably promote the development of rice research.