遥感影像光谱特征，植被指数特征，纹理特征，树高特征提取随机森林分类代码

以下是基于 Python 的遥感影像分类代码示例，使用了随机森林算法，提取了光谱特征、植被指数特征、纹理特征和树高特征。

# 导入所需的库
import numpy as np
import pandas as pd
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
from skimage.feature import greycomatrix, greycoprops
from osgeo import gdal

# 读取图像
image = gdal.Open('image.tif')
bands_data = []
for i in range(image.RasterCount):
    bands_data.append(image.GetRasterBand(i + 1).ReadAsArray())
bands_data = np.array(bands_data)

# 定义函数，用于计算纹理特征
def compute_texture_feature(band_data):
    # 计算灰度共生矩阵
    glcm = greycomatrix(band_data, [1], [0, np.pi/4, np.pi/2, 3*np.pi/4], levels=256, normed=True, symmetric=True)
    # 计算纹理特征
    contrast = greycoprops(glcm, 'contrast').ravel()
    dissimilarity = greycoprops(glcm, 'dissimilarity').ravel()
    homogeneity = greycoprops(glcm, 'homogeneity').ravel()
    energy = greycoprops(glcm, 'energy').ravel()
    correlation = greycoprops(glcm, 'correlation').ravel()
    asm = greycoprops(glcm, 'ASM').ravel()
    return np.hstack([contrast, dissimilarity, homogeneity, energy, correlation, asm])

# 定义函数，用于提取特征
def extract_features(bands_data):
    # 计算光谱特征
    spectral_features = np.transpose(bands_data, [1, 2, 0]).reshape(-1, bands_data.shape[0])
    # 计算植被指数特征
    ndvi = (bands_data[3] - bands_data[2]) / (bands_data[3] + bands_data[2])
    ndvi_features = ndvi.reshape(-1, 1)
    # 计算纹理特征
    texture_features = compute_texture_feature(bands_data[3])
    # 计算树高特征
    tree_height = bands_data[4]
    tree_height_features = tree_height.reshape(-1, 1)
    # 合并所有特征
    features = np.hstack([spectral_features, ndvi_features, texture_features, tree_height_features])
    return features

# 提取特征
features = extract_features(bands_data)

# 读取样本数据和标签
train_data = pd.read_csv('train_data.csv')
train_labels = train_data['label'].values

# 训练随机森林分类器
clf = RandomForestClassifier(n_estimators=200, max_depth=10, random_state=0)
clf.fit(features[train_data.index], train_labels)

# 预测测试数据并计算准确率
test_data = pd.read_csv('test_data.csv')
test_labels = test_data['label'].values
test_features = extract_features(bands_data)
test_pred = clf.predict(test_features[test_data.index])
accuracy = accuracy_score(test_labels, test_pred)
print('Accuracy:', accuracy)

其中，`image.tif` 是待分类的遥感影像，`train_data.csv` 和 `test_data.csv` 是训练集和测试集的样本数据和标签，示例代码中使用了光谱数据的第 1~4 波段作为光谱特征，第 5 波段作为树高特征。你可以根据实际情况进行修改。