Ambient-awareness in agriculture: 3-D structure and reasoning in the crop field (AACr3).

This dataset was created to enable research on semantic SLAM in agriculture, using a vineyard as a case study. The research goal is the automated recovery of geo-referenced 3-D structure of crop fields and the detection and classification of objects of interest, such as trunks, leaves, and fruits. The instances are readings of different sensors: a color stereo camera, a color 4K camera, a 16-channels LIDAR, a GNSS sensor with ground-base correction (Real-Time Kinematics), and IMUs. All sensors were assembled in a system mounted on a push cart and moved along the rows of two vineyards plots. The plots presented two different growing management systems: vertical shoot positioning growing (VSP) and Y-trellis (YT). This work was funded by the São Paulo Research Foundation (FAPESP) and IBM under grant 17/19282-7, Ambient-awareness in agriculture: 3-D structure and reasoning in the crop field (AACr3). For a complete description of this dataset, see README.md.