Dataset of tomato inside greenhouses for object detection in Pascal VOC

The development of robotic solutions for agriculture requires advanced perception capabilities that can work reliably in any crop stage. ROBOCARE is a project from CRIIS - INESC TEC, which focus on the development of robotics solutions for tomatoes' monitoring and harvesting inside greenhouses. Therefore, this project requires a reliable computer vision system to classify and localise tomatoes in the tomatoes' plants in the greenhouses. The current dataset was collected using a stereo camera (ZED camera) assembled in a static manipulator over a robotic platform to increase the similarity with the acquired data by the robot during its operation. The dataset was manually labeled using a dedicated annotation tool and is distributed under its full-size images (1280x720 px) or splitted images with 300x300 px considering an overlapping of 20% between adjacent images. This dataset is part of the journal paper submitted to Sensors and is distributted to allow the reproducibility of the performed work.