Date Fruit Dataset for Automated Harvesting and Visual Yield Estimation

The date fruit dataset was created to address the requirements of many applications in the pre-harvesting and harvesting stages. The two most important applications among them are automatic harvesting and visual yield estimation. The first dataset consists of 8079 images of more than 350 date bunches captured from 29 date palms. The date bunches belong to five date types: Naboot Saif, Khalas, Barhi, Meneifi, and Sullaj. The pictures of date bunches were captured using a color camera in six imaging sessions. The imaging sessions covered all date maturity stages: immature, Khalal, Rutab, and Tamar. The dataset is provided with a large degree of variations that represent the challenges occurs in a natural environment and date fruit orchards. This variation in images includes different angels and scales, different daylight conditions having poor illumination images, and date fruits covered by bags. The dataset was fully labeled according to type, maturity, and harvesting decision. We can use this dataset in many applications including fruit detection, segmentation, classification, maturity analysis, and automatic harvesting. The second dataset contains images, videos, and weight measurements to help in many applications such as yield estimation. In this dataset, we marked date bunches for selected palms, recorded 360° video for each palm, and measured their data (height, trunk circumference, total yield, number of bunches, and weight of bunches). We also captured images of each bunch from different angles before harvesting and on a graph paper after harvesting. Both datasets have been arranged with a coding scheme to simplify referring, linking, and facilitating future extensions of the dataset.本椰枣果实（date fruit）数据集旨在满足收获前与收获阶段诸多应用场景的需求。其中两类核心应用为自动采收与可视化产量预估。第一部分数据集包含8079幅图像，涵盖从29棵椰枣树（date palm）上采集的350余串椰枣果串（date bunch）。这些果串共涵盖5个椰枣品种：纳布特赛夫（Naboot Saif）、哈拉丝（Khalas）、巴尔希（Barhi）、米奈菲（Meneifi）与苏拉吉（Sullaj）。所有果串图像均通过彩色相机在6次成像会话中采集完成，本次成像覆盖椰枣全部成熟周期：未成熟期（immature）、卡拉尔期（Khalal）、鲁塔布期（Rutab）与塔马尔期（Tamar）。本数据集包含大量真实自然环境与椰枣果园中常见的各类挑战场景，图像差异涵盖不同拍摄角度与尺度、光照不佳的不同日光条件场景，以及套袋的椰枣果实。本数据集已按照品种、成熟度与采收决策完成全标注，可应用于诸多场景，包括果实检测、图像分割、品类分类、成熟度分析以及自动采收。第二部分数据集包含图像、视频与重量测量数据，可支撑产量预估等多类应用。在该数据集中，研究人员对选定椰枣树的果串进行了标记，为每棵树录制了360°视频，并测量了相关数据（树高、树干周长、总产量、果串数量与单串重量）。同时，我们在采收前从不同角度采集了每串果串的图像，并在采收后在坐标纸上拍摄了果串图像。两套数据集均采用编码方案进行整理，以简化引用、关联操作，并便于未来的数据集扩展工作。