Tutorials
Last updated
Last updated
This section provides a collection of tutorials on how to use the GRYFN Plot Extraction Tool.
ACRE-1 is a maize field in four row plots (Figure 1). The field consisted of 24 rows and 15 ranges. The segment length is 5 meters.
List of data:
LiDAR point cloud (LAS)
LiDAR Digital Surface Model (DSM) (TIFF)
RGB orthomosaic (TIFF)
Hyperspectral orthomosaic (TIFF)
Field boundary (JSON)
A basic plot extraction process includes three steps: new project (or open existing project), plot extraction, and labeling.
Creating a new project to start.
Specify output file prefix and default location.
Add data: LiDAR point cloud, DSM, RGB/hyperspectral orthomosaic.
The tool reads the header of each dataset and checks if the Coordinate Reference System (CRS) is consistent. If the CRS is not found in the header (usually for LAS), it asks the user to input the EPSG code manually. The EPSG code for West Lafayette, IN is 32616 (WGS 84 / UTM zone 16N).
Define area of interest (field boundary): Load a base map and a field boundary JSON file (or create a new one). Drag the vertices to adjust the area of interest in the graphic view.
Recommend using low resolution RGB for base map.
Include additional geometries if needed.
Add trial information. The segment length and width will be used to set default plot extraction parameters.
Create project. The information will be saved as a YAML file. The user can load the project YAML file using Open Project in the future.
See also New Project
Detecting plots in the field.
Select and load one dataset.
For point cloud or DSM, we recommend using mid to late season data when plants have attained considerable height above ground level. For RGB orthomosaic or vegetation index, we recommend using early season data as it allows for clear distinction between individual plant rows.
The resolution (cell size) is closely tied to the detection precision. For fields with a nominal 30-inch (0.76 meters) row spacing, the recommended value for resolution falls between 5 to 10 centimeters.
Vegetation index will be evaluated when loading an RGB orthomosaic (3-band TIFF). We currently support: Green Leaf Index (GLI), Normalized Green-Red Difference Index (NGRDI), and Visible Atmospherically Resistant Index (VARI).
Estimate the rotation angle. The rotation angle between the North (or East) and planting direction is automatically estimated upon loading the dataset. The user can fine-tune the value if needed. Select row direction.
Detect rows/alleys. Specify the row detection option based on the actual field condition (refer to Plot Extraction). Adjust the parameters (minimum spacing row & alley) if needed. The user can fine-tune detection sensitivity after initial detection.
Edit the detections. The graphic view displays the detected row centers (or gaps) and alleys as red lines. The user can select and move a detection to adjust its location. In addition, the Edit function allows for adding/deleting items and updating IDs.
Export plots.
Figure 4 shows the as-designed plots in QGIS. Note that the default labeling start from the top left corner.
See also Plot Extraction
Creating range, row, and segment IDs.
Select and load a base map. Load the plots (and corresponding centroids and centerlines, if any) to be labeled.
Select origin, i.e. location of the first range and row. Set the parameters (see examples below).
Row start ID: 2 Step: 2
Row IDs: 2, 4, 6, 8, ...
Segment per plot: 2 Consider planter's direction: No No. planter's row: 4
Segment IDs: 1, 2, 1, 2, 1, 2, 1, 2, ...
Segment per plot: 2 Consider planter's direction: Yes No. planter's row: 4
Segment IDs: 1, 2, 1, 2, 2, 1, 2,
Label and export plot.
Figure 6 shows the labeled as-designed plots in QGIS.
See also Labeling
