croplands/polygonize.py

import numpy as np
import rasterio
import geopandas as gpd
import cv2
from shapely.geometry import shape,Polygon
from shapely.validation import make_valid
from collections import defaultdict


def onehot(components: np.ndarray) -> np.ndarray:
    oh = np.zeros((components.max() + 1, *components.shape), dtype=np.uint8)
    for i in range(oh.shape[0]):
        oh[i][components == i] = 1
    if 0 in np.unique(components):
        oh = oh[1:]
    return oh

def apply_transform(polygon: Polygon, transform: rasterio.Affine) -> Polygon:
    return Polygon([transform * c for c in polygon.exterior.coords])

def polygonize_raster(
    data_img:np.ndarray,
    raster: np.ndarray,
    tolerance: float,
    transform: rasterio.transform.Affine = None,
    crs = None,
    pixel_threshold: int = 25,
) -> gpd.GeoDataFrame:
    data = defaultdict(list)
    onehot_raster = onehot(raster)
    if onehot_raster.shape[0] == 0:
        return None
    for i in range(onehot_raster.shape[0]):
        mask = onehot_raster[i]
        if mask.sum() < pixel_threshold:
            continue

        contours, _ = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
        for contour in contours:
            perimeter = cv2.arcLength(contour, True)
            area = cv2.contourArea(contour)
            approx = cv2.approxPolyDP(contour, tolerance * perimeter, True)
            contour = approx.squeeze()
            if contour.shape[0] < 3:
                continue
            poly = shape({"type": "Polygon", "coordinates": [contour]})
            if transform is not None:
                poly = apply_transform(poly, transform)

            most_freq_class = np.argmax(np.bincount(data_img[onehot_raster[i].astype(bool)]))
            data["geometry"].append(poly)
            data["component"].append(i)
            data["class"].append(most_freq_class)
            data["area"].append(area)

    gdf = gpd.GeoDataFrame(data, crs=crs)
    gdf["geometry"] = gdf["geometry"].buffer(9.0)
    gdf["geometry"] = gdf["geometry"].apply(make_valid)
    return gdf