Photometric stereo test data set

A test data set for extracting trait data from Arabidopsis thaliana plants using PS-Plant software. The data set included raw and processed data, and a pre-trained Mask R-CNN model. This data set forms part of work done in the BBSRC Tools and Resources Development project BB/N02334X/1. All PS-Plant software packages are available for download (https://bit.ly/2EFOk0O). Installation instructions are available on YouTube: https://youtu.be/7q6ZnJvhx6I. The PS-Plant software pipeline has seven steps: (1) PS data acquisition, (2) adaptive light source generation, (3) raw PS data processing, (4) rosette and (5) individual leaf mask generation, and (6) leaf instance tracking. The generation of compiled results (i.e. plant trait data) can be done at steps 4 and 6 to obtain rosette- and leaf-level data, respectively. Most PCs with onboard GPUs can be used to acquire and process PS data (steps 1-4 and 6). We recommend a PC with at least an Intel Core i3 equivalent CPU, and 4 GB of RAM. A GPU with a minimum Compute Unified Device Architecture (CUDA) capability of 3.5 (https://bit.ly/1Jnzfz8) is required for the generation of leaf masks (step 5) using the Leaf Segmentation GUI.基于PS-Plant软件提取拟南芥（Arabidopsis thaliana）性状数据的测试数据集。本数据集包含原始数据、处理后数据以及一个预训练的Mask R-CNN模型。本数据集是BBSRC工具与资源开发项目BB/N02334X/1相关研究工作的组成部分。所有PS-Plant软件安装包均可通过链接https://bit.ly/2EFOk0O下载获取，软件安装教程可在YouTube平台查看，链接为https://youtu.be/7q6ZnJvhx6I。PS-Plant软件流程共包含七个步骤：(1) PS数据采集，(2) 自适应光源生成，(3) 原始PS数据处理，(4) 莲座丛掩码生成，(5) 单叶掩码生成，以及(6) 叶片实例追踪。可在步骤4和步骤6分别生成整合结果（即植物性状数据），以获取莲座丛层级与叶片层级的性状数据。多数搭载集成GPU的个人计算机即可完成PS数据的采集与处理（对应步骤1-4及步骤6）。我们推荐使用至少搭载英特尔酷睿i3级别的CPU、内存不少于4GB的个人计算机。若需通过叶片分割图形用户界面（Leaf Segmentation GUI）完成叶片掩码生成（步骤5），则需使用至少具备3.5版本统一计算设备架构（Compute Unified Device Architecture，简称CUDA）算力的GPU，相关说明可参考链接https://bit.ly/1Jnzfz8。