论文部分内容阅读
植物表型自动化检测技术在农业研究和作物育种的过程中发挥了重要作用,但目前受限于二维技术三维特征很难被提取。株型是影响多分蘖作物产量的重要表型特征之一,它包括分蘖数、分蘖角、和茎粗等参数。传统方法中获取这些特征参数需要大量的人工测量,而人工测量具有耗时,主观性强,不准确等缺陷,因此用人工的方法进行大批量的表型分析是不现实的。为了使作物育种研究中株型参数提取实现自动化,提出一种用于高通量植株株型性状参数获取的快速三维重建方法,为了提高重建效率,研究中使用了图形处理单元(GPU)并行处理技术,在统一计算设备架构(CUDA)下进行重建的并行计算,使单株重建时间缩减到10秒左右,适合使用于高通量表型检测平台。
Plant phenotypic automated detection technology plays an important role in agricultural research and crop breeding, but the three-dimensional features limited by two-dimensional technology are difficult to be extracted at present. Plant type is one of the important phenotypic characteristics that affect the yield of multi-tiller crops. It includes parameters such as tiller number, tiller angle and stem diameter. The traditional method requires a lot of manual measurement to obtain these characteristic parameters. However, the manual measurement has the defects of time-consuming, subjective and inaccurate. Therefore, it is unrealistic to use artificial methods for large-scale phenotypic analysis. In order to automate the extraction of plant type parameters in crop breeding research, a rapid 3D reconstruction method for plant type trait parameters acquisition in high-throughput plants was proposed. In order to improve the reconstruction efficiency, a parallel processing unit (GPU) Technology, the parallel computing of reconstruction under the Unified Computing Architecture (CUDA) reduces the time to rebuild a plant to about 10 seconds, making it suitable for use in high-throughput phenotyping platforms.