Preparing Your Data¶

Getting Data Ready for crop-row-connector¶

This guide walks through preparing real field data for processing with crop-row-connector.

Data Sources¶

Your data comes from:

Crop Row Detector (prerequisite tool)
- Processes orthomosaic imagery
- Outputs row geometry data
Orthomosaic Image (optional for visualization)
- GeoTIFF format (.tif)
- Used with QGIS for verification
Field Survey Data (if available)
- Ground control points
- GPS coordinates
- Can improve accuracy

Required Input: Row Information CSV¶

This file contains detected crop row segments from Crop Row Detector.

File Structure¶

Required Columns (in this order):

tile	x_position	angle	row	x_start	y_start	x_end	y_end	x_mid	y_mid
1	0	1.5708	0	100.5	50.2	100.5	150.3	100.5	100.25
1	0	1.5708	1	105.2	49.8	105.2	149.9	105.2	99.85
2	1	1.5708	0	205.1	51.0	205.1	151.2	205.1	101.1

Note

This file is read by column position, not by header name — the column order must be exactly as shown, but the header text itself does not matter.

Column Definitions:

tile: Integer ID unique within this field (typically sequential: 1, 2, 3, …)
x_position: Tile X coordinate in the field grid
y_position: Tile Y coordinate in the field grid
angle: Crop row direction in radians (typically 0 to 2π)
row: Row number within this tile (unique per tile)
x_start: X coordinate of row segment start (in meters or field units)
y_start: Y coordinate of row segment start
x_end: X coordinate of row segment end
y_end: Y coordinate of row segment end
x_mid: X coordinate of row segment midpoint
y_mid: Y coordinate of row segment midpoint

Required Input: Vegetation Points CSV¶

This file contains individual vegetation measurement points from the orthomosaic.

File Structure¶

Required Columns (in this order):

tile	row	x	y	vegetation
1	0	100.5	50.2	200
1	0	100.5	50.5	195
1	0	100.5	50.8	198
1	1	105.2	49.8	212

Warning

The tile and row columns are looked up by name, so the header must use exactly these column names. The remaining columns are read by position.

Column Definitions:

tile: Tile number (must match tile numbers from row information)
row: Row number within tile (must match the row column in row information)
x: X coordinate of point
y: Y coordinate of point
vegetation: Vegetation intensity (0-255 grayscale)
- 0 = bare soil
- 255 = dense healthy vegetation
- 127 = intermediate

Coordinate System Considerations¶

Understanding Your Coordinates¶

Your data can use different coordinate systems:

Local Grid: Tile (x, y) in grid coordinates (0,0 at top-left)
Meters: Absolute position in meters from field origin
GPS: Latitude/longitude (must be converted to meters)
Pixels: Image coordinates (must be georeferenced)

The crop-row-connector works with any consistent system as long as coordinates are in the same system.

Setting Up Your Project Directory¶

Recommended File Organization¶

my_field/
├── input/
│   ├── row_information.csv      (from Crop Row Detector)
│   ├── vegetation_points.csv    (from Crop Row Detector or other source)
│   └── orthomosaic.tif          (optional, for verification)
└── output/
    ├── connected_crop_rows.csv
    ├── vegetation_points.csv
    ├── healthy_segments.csv
    └── unhealthy_segments.csv

Quick Checklist¶

If processing fails ensure the following before processing your field data again:

[ ] Row information CSV has all required columns
[ ] Vegetation points CSV has all required columns
[ ] No duplicate rows in either CSV
[ ] No missing values (NaN) in numeric columns
[ ] Vegetation values in range 0-255
[ ] Angles in valid range (0-2π or -π to π)
[ ] Row numbers consistent with detected rows
[ ] Coordinates are in a consistent system
[ ] No special characters in file paths that could cause issues