================================================================================ Ladybird Cobbitty 2017 Brassica Dataset ================================================================================ Introduction ================================================================================ This dataset contains weekly scans of cauliflower and broccoli covering a ten week growth cycle from transplant to harvest. The dataset includes ground-truth, physical characteristics of the crop; environmental data collected by a weather station and a soil-sensor network; and scans of the crop performed by Ladybird, an autonomous agricultural robot. The scans include stereo colour, thermal and hyperspectral imagery. The crop were planted at Lansdowne Farm, a University of Sydney agricultural research and teaching facility. Lansdowne Farm is located in Cobbitty, a suburb 70km south-west of Sydney in New South Wales (NSW), Australia. Four 80 metre raised crop beds were prepared with a North-South orientation. Approximately 144 Brassica were planted in each bed. Cauliflower were planted in the first and third bed (from west to east). Broccoli were planted in the second and fourth beds. Sources of data ================================================================================ Three sources of data were collected: Manual measurements ---------------------------------------------------------------------------- Physical plant measurements were collected manually by measuring physical attributes of the plants directly. This includes non-destructive measurements such as plant greenness and height. Destructive measurements such as fresh and dry plant mass were also collected. The manual measurements are provided as comma-separated values (CSV) files. Insitu sensors ---------------------------------------------------------------------------- To monitor atmospheric conditions, a Davis Wireless Vantage Pro2 (6327AU) weather station was installed to log data across the entire ten week period. Amongst other variables, temperature, wind speed, humidity, pressure and rain-fall were recorded. The weather station data are provided in a CSV file. Soil sensors were installed in 16 locations across the vegetable beds. At each location data was collected at a 10cm and 30cm depth. In total, 32 Decagon GS3 sensors were logged. The soil-sensor network was configured to log temperature, electrical conductivity, dielectric and volumetric water content. The soil sensor network data are provided in a CSV file. Autonomous measurements ---------------------------------------------------------------------------- High-resolution, high-coverage proximal measurements were gathered autonomously by Ladybird - a multi-purpose agricultural robotics platform designed and built at the Australian Centre for Field Robotics, The University of Sydney. Ladybird has the following capabilities: - Localisation is provided via a NovAtel SPAN global positioning and inertial navigation system (GPS/INS). Real-time kinematic (RTK) corrections are provided to the system via a 4G broadband cellular connection. Navigation data are provided in CSV files. - Two Point Grey Grasshopper3 12MP GS3-U3-120S6C-C cameras are arranged in a stereo configuration with a baseline of 150mm. Each camera has a Schneider 10mm f/1.9 C-Mount Cinegon lens providing a cross-track field of view of 66.06 degrees. At the expected imaging height, this equates to approximately 987mm. The images are provided as timestamped PNG files in separate archives for the left and right cameras in the stereo pair. - A Resonon Pika XC-2 hyperspectral camera collecting line-scan imagery at 100 Hz. The camera is fitted with a Schneider 23mm f/1.4 Xenoplan lens. The focal length provides a cross-track field of view of 21.42 degrees. At the expected imaging height and angle, the field of view is approximately 1108mm. The hyperspectral line scans are provided as timestamped PNG files. - A Xenics Gobi-640 thermal camera collecting images at 10 Hz. The camera has a permanently installed 10mm lens providing a cross-track field of view of 56.84 degrees. At the expected imaging height, this covers approximately 807.26mm of the bed width. The images are provided as timestamped PNG files. Data Summary ================================================================================ Manual measurements and autonomous scans were performed weekly for ten weeks in October and November of 2017. The number of manual measurements collected is provided in the following table: | Week | Date (2017) | SPAD | Height & | Total fresh/dry | Head fresh/dry | | | | | Diameter | weight | weight | +------+---------------+--------+----------+-----------------+----------------+ | 1 | 12 October | - | - | - | - | | 2 | 19 October | 287 | - | - | - | | 3 | 25 October | 284 | 32 | - | - | | 4 | 1 November | 278 | 32 | 32 | - | | 5 | 8 November | 266 | 32 | 32 | - | | 6 | 15 November | 251 | 32 | 32 | - | | 7 | 22 November | 249 | 32 | 32 | 7 | | 8 | 29 November | 238 | 32 | 32 | 23 | | 9 | 7 December | 220 | 32 | 31 | 28 | | 10 | 14 December | 100 | 16 | 16 | 16 | The number of images collected by Ladybird is provided in the following table: | Week | Date (2017) | Stereo | Thermal | Hyperspectral | | | | images | images | line-scans | +------+---------------+--------+----------+-----------------+ | 1 | 12 October | 1758 | 33246 | 353313 | | 2 | 19 October | 1764 | 33771 | 353100 | | 3 | 25 October | 1744 | 33366 | 348312 | | 4 | 1 November | 1867 | 36152 | 376326 | | 5 | 8 November | 3580 | 34369 | 345367 | | 6 | 15 November | 3514 | 33274 | 353349 | | 7 | 22 November | 632 | 34210 | 354801 | | 8 | 29 November | 3578 | 34234 | 360050 | | 9 | 7 December | 3517 | 33846 | 352298 | | 10 | 14 December | 3411 | 16444 | 171151 | Reference ================================================================================ For a more detailed description of the dataset, or if any data is used in an academic publication, please refer to or cite: @article{Bender2020, author = {Bender, Asher and Whelan, Brett and Sukkarieh, Salah}, title = {A high-resolution, multimodal data set for agricultural robotics: A Ladybird's-eye view of Brassica}, journal = {Journal of Field Robotics}, volume = {37}, number = {1}, pages = {73-96}, doi = {10.1002/rob.21877}, url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/rob.21877}, eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1002/rob.21877}, year = {2020} } License ================================================================================ Ladybird Cobbitty 2017 Brassica Dataset (c) by Asher Bender, Brett Whelan, Salah Sukkarieh Ladybird Cobbitty 2017 Brassica Dataset is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. A copy of the license is included in LICENSE.txt, if it cannot be accessed, see .