監督分類：用隨機森林做遙感影像像素級分類

像素級分類有點耗時間，能力有限，在此先提供一個初版
這裏導入了傳統決策樹、隨機森林、極端決策樹，大家可以都試下
下面是跑了4波段遙感影像的代碼，最好選個小圖，我這裏逐像素分類很慢，還有很多需要改進的地方，希望大家可以在評論下給我點建議，指導一下
這裏是對***植被***進行提取

# -*- coding: utf-8 -*-
import os, sys, time
import gdal
from osgeo import ogr
from osgeo import gdal
from osgeo import gdal_array as ga
from gdalconst import *
from skimage import morphology,filters
import numpy as np
from numba import jit, vectorize, int64
import warnings 
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import ExtraTreesClassifier

#讀遙感影像
def read_img(filename):
    dataset=gdal.Open(filename)

    im_width = dataset.RasterXSize
    im_height = dataset.RasterYSize

    im_geotrans = dataset.GetGeoTransform()
    im_proj = dataset.GetProjection()
    im_data = dataset.ReadAsArray(0,0,im_width,im_height)

    del dataset 
    return im_proj,im_geotrans,im_width, im_height,im_data

#寫出圖像
def write_img(filename, im_proj, im_geotrans, im_data):
    if 'int8' in im_data.dtype.name:
        datatype = gdal.GDT_Byte
    elif 'int16' in im_data.dtype.name:
        datatype = gdal.GDT_UInt16
    else:
        datatype = gdal.GDT_Float32

    if len(im_data.shape) == 3:
        im_bands, im_height, im_width = im_data.shape
    else:
        im_bands, (im_height, im_width) = 1,im_data.shape 

    driver = gdal.GetDriverByName("GTiff")
    dataset = driver.Create(filename, im_width, im_height, im_bands, datatype)

    dataset.SetGeoTransform(im_geotrans)
    dataset.SetProjection(im_proj)

    if im_bands == 1:
        dataset.GetRasterBand(1).WriteArray(im_data)
    else:
        for i in range(im_bands):
            dataset.GetRasterBand(i+1).WriteArray(im_data[i])

    del dataset

#根據矢量點獲取點對應的像素值，並把點先放入列表中
def getPixels(shp, img):
    driver = ogr.GetDriverByName('ESRI Shapefile')
    ds = driver.Open(shp, 0)
    if ds is None:
        print('Could not open ' + shp)
        sys.exit(1)

    layer = ds.GetLayer()

    xValues = []
    yValues = []
    feature = layer.GetNextFeature()
    while feature:
        geometry = feature.GetGeometryRef()
        x = geometry.GetX()
        y = geometry.GetY()
        xValues.append(x)
        yValues.append(y)
        feature = layer.GetNextFeature()

    gdal.AllRegister()

    ds = gdal.Open(img, GA_ReadOnly)
    if ds is None:
        print('Could not open image')
        sys.exit(1)

    rows = ds.RasterYSize
    cols = ds.RasterXSize
    bands = ds.RasterCount

    transform = ds.GetGeoTransform()
    xOrigin = transform[0]
    yOrigin = transform[3]
    pixelWidth = transform[1]
    pixelHeight = transform[5]

    values = []
    for i in range(len(xValues)):
        x = xValues[i]
        y = yValues[i]

        xOffset = int((x - xOrigin) / pixelWidth)
        yOffset = int((y - yOrigin) / pixelHeight)

        s = str(int(x)) + ' ' + str(int(y)) + ' ' + str(xOffset) + ' ' + str(yOffset) + ' '

        dt = ds.ReadAsArray(xOffset, yOffset, 1, 1)
        values.append(dt.flatten())
    return values

if __name__ == "__main__":
    img_path = "E:/20200210/forest/gf2/dys_gf2.tif"  #原始大圖，在這上面選點
    img_path2 = "E:/20200210/forest/gf2/dys_gf2_test.tif"  #測試小圖，測試用
    shp_false = "E:/20200210/forest/point/1.shp" #負樣本，豐富一點
    shp_true = "E:/20200210/forest/point/2.shp"  #正樣本，越多越好，不要和負樣本混淆
    temp_path = "E:/20200210/forest/temp/"  #存放臨時文件
    
    point_false = getPixels(shp_false, img_path)
    num1 = len(point_false)
    lab_false = np.zeros((num1))

    point_true = getPixels(shp_true, img_path)
    num2 = len(point_true)
    lab_true = np.ones((num2))

    data = point_false + point_true
    label = list(lab_false) + list(lab_true)
    data = np.array(data)
    label = np.array(label)

    clf = RandomForestClassifier(n_estimators=100, max_depth=2,random_state=0)
    clf.fit(data, label)
    # print(clf.feature_importances_)

    im_proj2, im_geotrans2, im_width2, im_height2, im_data2 = read_img(img_path2)
    seg = np.zeros((im_data2.shape[1],im_data2.shape[2]))
    for i in xrange(im_width2-1):
        for j in xrange(im_height2-1):
            point = im_data2[0:4,j,i]
            point = np.expand_dims(point,0)
            seg[j,i] = clf.predict(point)[0]

    seg = np.int8(seg)
    seg_path = os.path.join(temp_path, 'random.tif')
    write_img(seg_path, im_proj2, im_geotrans2, seg)

測試圖像

輸出結果

說明：這裏我選點沒有完全覆蓋所有，所以可能會錯提一些，選點一定要好好選。這個結果很粗還需要加些後處理可能會好看點。

希望各位看官給我點加速的建議，這個跑太慢了。對於逐像素遍歷怎麼才能更快。或者不逐像素遍歷能不能實現。