使用 python 和 gdal 从 .geotiff 中的像素点查找 lat/long 坐标

Find lat/long coordinates from pixel point in .geotiff using python and gdal

我在 .geotiff 图像中有由像素坐标定义的点。

我正在尝试将这些像素坐标转换为经度和纬度。

起初我试图从我之前问过的逆向问题中改编给定的代码

def pixel_to_world(geo_matrix, x, y):
    ul_x   = geo_matrix[0]
    ul_y   = geo_matrix[3]
    x_dist = geo_matrix[1]
    y_dist = geo_matrix[5]
    _x = x * x_dist + ul_x
    _y = y * y_dist + ul_y
    return (_x, _y)

def build_transform_inverse(dataset, EPSG):
    source = osr.SpatialReference(wkt=dataset.GetProjection())
    target = osr.SpatialReference()
    target.ImportFromEPSG(4326)
    return osr.CoordinateTransformation(source, target)

def find_spatial_coordinate_from_pixel(dataset, transform, x, y):
    point = ogr.Geometry(ogr.wkbPoint)
    point.AddPoint(float(x), float(y))
    point.Transform(transform)
    return pixel_to_world(dataset.GetGeoTransform(), point.GetX(), point.GetY())

ds = gdal.Open(source_directory_path + filename)
_t = build_transform_inverse(ds, 4326)
coordinates = find_spatial_coordinate_from_pixel(ds, _t, point[0], point[1])

但我的最终坐标不正确,我得到的是 (-1528281.9351183572, 1065990.778732022) 而不是 (-13.728790283203121, 9.531686027524676)

ul_x = -1528281.943533814, 
ul_y = 1065990.7964650677, 
x_dist = 0.5970709765586435, 
y_dist = -0.5972870511184641

我也试过像这样的光栅方法:

import rasterio

with rasterio.open('path/to/file.tiff') as map_layer:
    pixels2coords = map_layer.xy(2679,2157)
    print(pixels2coords)

无法 return 合适的长纬度坐标对(它 returns (-1526993.7629018887, 1064390.365811596))。

我做错了什么?

您颠倒了像素、投影坐标和经纬度之间的转换顺序。你正在做的订单 pixels -> lat lngs -> projected 它应该是 pixels -> projected -> lat lngs.

以下应该有效

from osgeo import osr, ogr, gdal


def pixel_to_world(geo_matrix, x, y):
    ul_x = geo_matrix[0]
    ul_y = geo_matrix[3]
    x_dist = geo_matrix[1]
    y_dist = geo_matrix[5]
    _x = x * x_dist + ul_x
    _y = y * y_dist + ul_y
    return _x, _y


def build_transform_inverse(dataset, EPSG):
    source = osr.SpatialReference(wkt=dataset.GetProjection())
    target = osr.SpatialReference()
    target.ImportFromEPSG(EPSG)
    return osr.CoordinateTransformation(source, target)


def find_spatial_coordinate_from_pixel(dataset, transform, x, y):
    world_x, world_y = pixel_to_world(dataset.GetGeoTransform(), x, y)
    point = ogr.Geometry(ogr.wkbPoint)
    point.AddPoint(world_x, world_y)
    point.Transform(transform)
    return point.GetX(), point.GetY()

ds = gdal.Open(source_directory_path + filename)
_t = build_transform_inverse(ds, 4326)
coordinates = find_spatial_coordinate_from_pixel(ds, _t, point[0], point[1])
print(coordinates)

find_spatial_coordinate_from_pixel 中我颠倒了顺序,现在 pixel_to_world 在坐标转换之前首先被调用。