CombineCropRows

Qualified name: crop_row_connector.combine_crop_rows.CombineCropRows

class CombineCropRows(angle_tolerance, vegetation_threshold, min_unhealthy_vegetation_length, max_segment_length, distance_tolerance, *, output_path_connected_crop_rows=None, output_path_vegetation_points=None, output_path_healthy_vegetation_segments=None, output_path_unhealthy_vegetation_segments=None, max_workers=None)

Bases: object

Combine crop rows from different tiles into one line.

Parameters:

• angle_tolerance (float) – Maximum angle difference in degrees for rows to be considered for connection.

• vegetation_threshold (float) – Grayscale threshold (0-255) for classifying vegetation as healthy.

• min_unhealthy_vegetation_length (float) – Minimum spatial extent in meters for unhealthy vegetation segments to be recorded.

• max_segment_length (float) – Maximum length in meters for vegetation classification segments.

• output_path_connected_crop_rows (Path | None) – Path to save connected crop rows to CSV.

• output_path_vegetation_points (Path | None) – Path to save all vegetation points associated with connected rows.

• output_path_healthy_vegetation_segments (Path | None) – Path to save segments with healthy vegetation.

• output_path_unhealthy_vegetation_segments (Path | None) – Path to save segments with unhealthy vegetation.

• max_workers (int | None) – Maximum number of parallel workers for processing.

• distance_tolerance (float)

Methods

connect_2_tiles	Connect two tiles by linking crop rows if their angles are within the allowed tolerance.
connect_crop_rows	Execute the complete crop row connection pipeline.
connect_rows_in_tiles	Connect crop rows across neighboring tiles in the grid.
connected_crop_rows_to_csv	Write the connected crop rows to a csv file.
create_tile_grid	Create a 2D grid of tile numbers in their respective positions, with -1 for empty cells.
length_of_segments	Calculate total lengths of healthy and unhealthy vegetation segments.
load_csv	Load a CSV file containing row information into a NumPy array.
merge_all_points_in_all_crop_rows_remove	Merge all points from connected crop rows into a single DataFrame, removing overlapping points based on a distance tolerance.
separate_healthy_and_unhealthy_vegetation_segments	Define healthy and unhealthy vegetation segments and optionally write them to shapefiles.
separate_row_information_to_tile	Split row information into tile objects keyed by tile number.

connect_2_tiles(tile_1, tile_2)

Connect two tiles by linking crop rows if their angles are within the allowed tolerance.

Parameters:

• tile_1 (Tile) – The first tile object to connect.

• tile_2 (Tile) – The second tile object to connect.

Return type:

None

connect_crop_rows(path_row_information, path_points_in_rows)

Execute the complete crop row connection pipeline. And save desired output.

Parameters:

• path_row_information (Path) – Path to CSV with row information. Format: [tile_number, x_position, y_position, angle, row, x_start, y_start, x_end, y_end]

• path_points_in_rows (Path) – Path to CSV with vegetation points. Format: [tile_number, row_number, x, y, vegetation]

Return type:

None

connect_rows_in_tiles(grid, tiles, tqdm_disable=False)

Connect crop rows across neighboring tiles in the grid.

For each tile in the grid: - Connect to its right neighbor (if any) - Connect to its bottom neighbor (if any)

After all connections are made: - Add any uncombined rows to the connected crop rows

Parameters:

• grid (ndarray) – 2D array representing the tile grid. Each entry is a tile number, or -1 for empty cells.

• tiles (dict[int, Tile]) – Dictionary mapping tile numbers to tile objects.

• tqdm_disable (bool)

Return type:

None

connected_crop_rows_to_csv(connected_crop_rows)

Write the connected crop rows to a csv file.

Parameters:

connected_crop_rows (ndarray) – 2D array containing connected crop row information.

Return type:

None

create_tile_grid(row_information, tiles)

Create a 2D grid of tile numbers in their respective positions, with -1 for empty cells.

Notes

• X indexes columns, Y indexes rows (grid[y, x])

• The grid is padded by 1 row and 1 column to simplify neighbor connections.

Parameters:

• row_information (ndarray) – 2D array with row information including tile positions.

• tiles (dict[int, Tile]) – Dictionary mapping tile numbers to tile objects.

Returns:

2D array where each cell contains a tile number if present, otherwise -1.

Return type:

np.ndarray

length_of_segments(segments)

Calculate total lengths of healthy and unhealthy vegetation segments.

Notes

• Uses segment endpoints instead of point-to-point distances.

• Each segment is treated as a straight line.

Parameters:

segments (list[dict]) – Output from separate_healthy_and_unhealthy_vegetation_segments.

Returns:

(total_length, healthy_length, unhealthy_length)

Return type:

tuple[float, float, float]

load_csv(path)

Load a CSV file containing row information into a NumPy array.

Notes

• Each row corresponds to a crop row.

• Columns contain:

  [tile_number, x_position, y_position, angle, row, x_start, y_start, x_end, y_end, x_mid, y_mid]

Parameters:

path (Path)

Return type:

ndarray

merge_all_points_in_all_crop_rows_remove(connected_crop_rows, DF_vegetation_rows, row_information, tiles)

Merge all points from connected crop rows into a single DataFrame, removing overlapping points based on a distance tolerance.

Parameters:

• connected_crop_rows (ndarray) – 2D array containing connected crop row information.

• path_points_in_rows – Path to CSV containing raw vegetation points per row.

• row_information (ndarray) – Original row information array.

• tiles (dict[int, Tile]) – Dictionary mapping tile numbers to tile objects.

• DF_vegetation_rows (DataFrame)

Returns:

Merged and sorted crop rows including vegetation and duplicate information.

Return type:

pd.DataFrame

separate_healthy_and_unhealthy_vegetation_segments(DF_crop_rows)

Define healthy and unhealthy vegetation segments and optionally write them to shapefiles.

Notes

• Uses self.vegetation_threshold to separate healthy from unhealthy vegetation.

• Each segment is represented as a line between two points.

• Returns segments so they can be reused for further analysis (e.g. length calculation).

• X indexes columns, Y indexes rows (DF_crop_rows[‘x’], DF_crop_rows[‘y’]).

Parameters:

DF_crop_rows (DataFrame) – DataFrame containing at least [‘row’, ‘x’, ‘y’, ‘vegetation’].

Returns:

List of segment dictionaries: { “row_id”: int, “healthy”: list[[[x1, y1], [x2, y2]]], “unhealthy”: list[[[x1, y1], [x2, y2]]] }

Return type:

list[dict]

separate_row_information_to_tile(row_information)

Split row information into tile objects keyed by tile number.

Parameters:

row_information (ndarray) – A NumPy array containing row information for all tiles. Each row corresponds to a crop row and contains the following columns: [tile_number, x_position, y_position, angle, row, x_start, y_start, x_end, y_end]

Returns:

Dictionary mapping each tile number to a tile object.

Return type:

dict[int, Tile]