如何在python中将numpy列表转换为geoTIFF？

Question

我还没有找到这个简单问题的答案。请帮帮忙。如何将Qcal (数字列表)转换为TIFF图像？我发现的一切都不管用。

import math
import numpy as np
from osgeo import gdal

substr1 = 'RADIANCE_MULT_BAND_10'
substr2 = 'RADIANCE_ADD_BAND_10'
substr3 = 'K1_CONSTANT_BAND_10'
substr4 = 'K2_CONSTANT_BAND_10'

RADIANCE_MULT_BAND_10 = 1
RADIANCE_ADD_BAND_10 = 1
K1_CONSTANT_BAND_10 = 1
K2_CONSTANT_BAND_10 = 1

with open('LC08_L1TP_180028_20170623_20170630_01_T1_MTL.txt') as file:
    for line in file:
        if substr1 in line:
            startIndex = line.find('=')
            RADIANCE_MULT_BAND_10 = float((line[startIndex+2:]))
        if substr2 in line:
            startIndex = line.find('=')
            RADIANCE_ADD_BAND_10 = float((line[startIndex + 2:]))
        if substr3 in line:
            startIndex = line.find('=')
            K1_CONSTANT_BAND_10 = float((line[startIndex + 2:]))
        if substr4 in line:
            startIndex = line.find('=')
            K2_CONSTANT_BAND_10 = float((line[startIndex + 2:]))

ds = gdal.Open("B10.tif")
Qcal = np.array(ds.GetRasterBand(1).ReadAsArray()) # Quantized and calibrated standard product pixel values (DN)
for i in range(Qcal.shape[0]):
   for j in range(Qcal.shape[1]):
       Qcal[i][j] = RADIANCE_MULT_BAND_10 * Qcal[i][j] + RADIANCE_ADD_BAND_10
       Qcal[i][j] = K2_CONSTANT_BAND_10 / math.log1p(K1_CONSTANT_BAND_10/Qcal+1)

Answer 1

您要修改现有的映像吗？

如果是这样的话，您应该使用更新选项打开它，如下所示

gdal.Open("B10.tif", gdal.GA_Update)

接下来，就像正在做的那样，对np数组进行修改。实际上，您只是在编辑numpy数组Qcal，而不是tif中的实际光栅带。

现在，要将修改保存到相同的带中，可以执行以下操作

ds.GetRasterBand(1).WriteArray(Qcal)

这是将更新的Qcal数组写入tif的栅格带。

如果您想要保存为一个新的图像

您可以创建现有映像的副本，并将Qcal数组保存到其中，如下所示

driver = gdal.GetDriverByName('Gtiff')
dst_ds = driver.CreateCopy("example.tif", ds, 1)
dst_ds.GetRasterBand(1).WriteArray(Qcal)