684 lines
33 KiB
Python
684 lines
33 KiB
Python
# -*- coding: UTF-8 -*-
|
||
"""
|
||
@Project ：microproduct
|
||
@File ：LandCoverMain.py
|
||
@Function ：实现对数据地表覆盖的训练和分类识别
|
||
@Author ：SHJ
|
||
@Date ：2022/11/23
|
||
@Version ：2.0.0
|
||
"""
|
||
import datetime
|
||
import glob
|
||
import logging
|
||
import os
|
||
import shutil
|
||
import sys
|
||
from PIL import Image
|
||
import numpy as np
|
||
import cv2
|
||
from osgeo import gdal, osr
|
||
import multiprocessing
|
||
import pyproj._compat
|
||
# 导入PreProcess模块要在其他包含gdal库的模块前面，不然剪切影像会报错，详见https://blog.csdn.net/u014656611/article/details/106450006
|
||
from tool.algorithm.algtools.PreProcess import PreProcess as pp
|
||
from LandCoverAuxData import LandCoverMeasCsv
|
||
from tool.algorithm.polsarpro.AHVToPolsarpro import AHVToPolsarpro
|
||
from tool.algorithm.image.ImageHandle import ImageHandler
|
||
from tool.algorithm.algtools.ROIAlg import ROIAlg as alg
|
||
from tool.algorithm.xml.AlgXmlHandle import ManageAlgXML, CheckSource, InitPara
|
||
from tool.algorithm.block.blockprocess import BlockProcess
|
||
from tool.algorithm.algtools.logHandler import LogHandler
|
||
from tool.algorithm.polsarpro.pspCloudePottierDecomposition import PspCloudePottierDecomposition
|
||
from tool.algorithm.polsarpro.pspFreemanDecomposition import PspFreemanDecomposition
|
||
from tool.algorithm.polsarpro.pspLeeRefinedFilterT3 import LeeRefinedFilterT3
|
||
from tool.algorithm.polsarpro.pspYamaguchiDecomposition import PspYamaguchiDecomposition
|
||
from tool.algorithm.polsarpro.pspTouziDecomposition import PspTouziDecomposition
|
||
from tool.algorithm.polsarpro.bin2tif import write_bin_to_tif
|
||
from tool.algorithm.xml.CreatMetafile import CreateMetafile
|
||
from tool.algorithm.algtools.MetaDataHandler import Calibration
|
||
from tool.algorithm.xml.CreateMetaDict import CreateMetaDict, CreateProductXml
|
||
from tool.algorithm.xml.AnalysisXml import DictXml, xml_extend
|
||
from LandCoverXmlInfo import CreateDict, CreateStadardXmlFile
|
||
from tool.config.ConfigeHandle import Config as cf
|
||
from tool.file.fileHandle import fileHandle
|
||
from tool.algorithm.transforml1a.transHandle import TransImgL1A
|
||
from tool.algorithm.ml.machineLearning import MachineLeaning as ml
|
||
from tool.csv.csvHandle import csvHandle
|
||
|
||
csvh = csvHandle()
|
||
|
||
if cf.get('debug') == 'True':
|
||
DEBUG = True
|
||
else:
|
||
DEBUG = False
|
||
EXE_NAME = cf.get('exe_name')
|
||
tar = r'-' + cf.get('tar')
|
||
productLevel = cf.get('productLevel')
|
||
LogHandler.init_log_handler('run_log\\' + EXE_NAME)
|
||
logger = logging.getLogger("mylog")
|
||
FILTER_SIZE = int(cf.get('filter_size'))
|
||
file =fileHandle(DEBUG)
|
||
MAX_TRAN_NUM = int(cf.get('max_tran__num_per_class'))
|
||
# env_str = os.path.split(os.path.realpath(__file__))[0]
|
||
env_str = os.path.dirname(os.path.abspath(sys.argv[0]))
|
||
os.environ['PROJ_LIB'] = os.getcwd()
|
||
class LandCoverMain:
|
||
"""
|
||
算法主函数
|
||
"""
|
||
def __init__(self, alg_xml_path):
|
||
self.alg_xml_path = alg_xml_path
|
||
self.imageHandler = ImageHandler()
|
||
self.__alg_xml_handler = ManageAlgXML(alg_xml_path)
|
||
self.__check_handler = CheckSource(self.__alg_xml_handler)
|
||
self.__workspace_path, self.__temp_path, self.__task_id, self.__out_para = None, None, None, None
|
||
self.__input_paras, self.__output_paras, self.__processing_paras,self.__preprocessed_paras = {},{}, {},{}
|
||
# 保存特征数据的名称
|
||
self.__feature_name_list = []
|
||
# xml参数字典
|
||
self.__parameters_dic = {}
|
||
# 特征表
|
||
self.__FeatureMap = {}
|
||
self.__TotalFeatureSize = 14 # 3+4+4+3
|
||
# 特征索引对应的文件名
|
||
self.___FeatureFileNameMap = {}
|
||
# 初始化特征参数
|
||
self.__FeatureParaInit()
|
||
# 输入特征参数
|
||
self.__FeatureInput = {}
|
||
|
||
# 参考影像路径
|
||
self.__ref_img_path = ''
|
||
# 宽/列数
|
||
self.__cols = 0
|
||
# 高/行数
|
||
self.__rows = 0
|
||
# 坐标系
|
||
self.__proj = ''
|
||
# 影像投影变换矩阵
|
||
self.__geo = [0, 0, 0, 0, 0, 0]
|
||
# 分块尺寸
|
||
self.__block_size = 0
|
||
self.processinfo = [1, 1, 1, 1]
|
||
|
||
def check_source(self):
|
||
"""
|
||
检查算法相关的配置文件，图像，辅助文件是否齐全
|
||
"""
|
||
env_str = os.getcwd()
|
||
logger.info("sysdir: %s", env_str)
|
||
|
||
self.__check_handler.check_alg_xml()
|
||
self.__check_handler.check_run_env()
|
||
|
||
# 检查景影像是否为全极化
|
||
self.__input_paras = self.__alg_xml_handler.get_input_paras()
|
||
checkFlag, self.__parameters_dic = self.__check_handler.check_input_paras(self.__input_paras)
|
||
self.__workspace_path = self.__alg_xml_handler.get_workspace_path()
|
||
self.__create_work_space()
|
||
self.__processing_paras = InitPara.init_processing_paras(self.__input_paras)
|
||
self.__processing_paras.update(self.get_tar_gz_inf(self.__processing_paras["sar_path0"]))
|
||
SrcImageName = os.path.split(self.__input_paras["AHV"]['ParaValue'])[1].split('.tar.gz')[0]
|
||
result_name = SrcImageName + tar + ".tar.gz"
|
||
self.__out_para = os.path.join(self.__workspace_path, EXE_NAME, 'Output', result_name)
|
||
self.__alg_xml_handler.write_out_para("LandCoverProduct", self.__out_para) #写入输出参数
|
||
logger.info('check_source success!')
|
||
return True
|
||
|
||
def get_tar_gz_inf(self, tar_gz_path):
|
||
para_dic = {}
|
||
name = os.path.split(tar_gz_path)[1].rstrip('.tar.gz')
|
||
file_dir = os.path.join(self.__workspace_preprocessing_path, name + '\\')
|
||
file.de_targz(tar_gz_path, file_dir)
|
||
# 元文件字典
|
||
para_dic.update(InitPara.get_meta_dic_new(InitPara.get_meta_paths(file_dir, name), name))
|
||
# tif路径字典
|
||
para_dic.update(InitPara.get_polarization_mode(InitPara.get_tif_paths(file_dir, name)))
|
||
parameter_path = os.path.join(file_dir, "orth_para.txt")
|
||
para_dic.update({"paraMeter": parameter_path})
|
||
return para_dic
|
||
|
||
def __create_work_space(self):
|
||
"""
|
||
删除原有工作区文件夹,创建新工作区文件夹
|
||
"""
|
||
self.__workspace_preprocessing_path = os.path.join(self.__workspace_path, EXE_NAME, "Temporary", "preprocessing\\") # self.__workspace_path + EXE_NAME + r"\Temporary\preprocessing""\\"
|
||
self.__workspace_preprocessed_path = os.path.join(self.__workspace_path, EXE_NAME, "Temporary", "preprocessed\\") # self.__workspace_path + EXE_NAME + r"\Temporary\preprocessed""\\"
|
||
self.__workspace_processing_path = os.path.join(self.__workspace_path, EXE_NAME, "Temporary", "processing\\") # self.__workspace_path + EXE_NAME + r"\Temporary\processing""\\"
|
||
self.__workspace_block_tif_path = os.path.join(self.__workspace_path, EXE_NAME, "Temporary", "blockTif\\") # self.__workspace_path + EXE_NAME + r"\Temporary\blockTif""\\"
|
||
self.__workspace_block_tif_processed_path = os.path.join(self.__workspace_path, EXE_NAME, "Temporary", "blockTifProcessed\\") # self.__workspace_path + EXE_NAME + r"\Temporary\blockTifProcessed""\\"
|
||
self.__feature_tif_dir = os.path.join(self.__workspace_processing_path, 'feature_tif') # self.__workspace_processing_path + 'feature_tif'
|
||
path_list = [self.__workspace_preprocessing_path, self.__workspace_preprocessed_path,
|
||
self.__workspace_processing_path, self.__workspace_block_tif_path,
|
||
self.__workspace_block_tif_processed_path, self.__feature_tif_dir]
|
||
file.creat_dirs(path_list)
|
||
logger.info('create new workspace success!')
|
||
|
||
|
||
def del_temp_workspace(self):
|
||
"""
|
||
临时工作区
|
||
"""
|
||
if DEBUG is True:
|
||
return
|
||
path = os.path.join(self.__workspace_path, EXE_NAME, "Temporary") # self.__workspace_path + EXE_NAME + r"\Temporary"
|
||
if os.path.exists(path):
|
||
file.del_folder(path)
|
||
|
||
def preprocess_handle(self):
|
||
"""
|
||
预处理
|
||
"""
|
||
|
||
para_names_geo = ['sim_ori']
|
||
self.__feature_name_list = para_names_geo
|
||
p = pp()
|
||
p.check_img_projection(self.__workspace_preprocessing_path, para_names_geo, self.__processing_paras)
|
||
# 计算roi
|
||
scopes = ()
|
||
scopes += (xml_extend(self.__processing_paras['META']).get_extend(),)
|
||
scopes += p.box2scope(self.__processing_paras['box'])
|
||
# 计算图像的轮廓,并求相交区域
|
||
intersect_shp_path = self.__workspace_preprocessing_path + 'IntersectPolygon.shp'
|
||
|
||
scopes_roi = p.cal_intersect_shp(intersect_shp_path, para_names_geo, self.__processing_paras, scopes)
|
||
# 裁剪
|
||
# 裁剪图像：裁剪微波图像，裁剪其他图像
|
||
cutted_img_paths = p.cut_imgs(self.__workspace_preprocessing_path, para_names_geo, self.__processing_paras,
|
||
intersect_shp_path)
|
||
self.__preprocessed_paras.update(cutted_img_paths)
|
||
|
||
para_names_l1a = ["HH", "VV", "HV", "VH"]
|
||
|
||
l1a_width = ImageHandler.get_img_width(self.__processing_paras['HH'])
|
||
l1a_height = ImageHandler.get_img_height(self.__processing_paras['HH'])
|
||
self._tr = TransImgL1A(self.__processing_paras['sim_ori'], scopes_roi,l1a_height,l1a_width) # 裁剪图像
|
||
|
||
for name in para_names_l1a:
|
||
out_path = os.path.join(self.__workspace_preprocessed_path, name + "_preprocessed.tif")
|
||
self._tr.cut_L1A(self.__processing_paras[name], out_path)
|
||
self.__preprocessed_paras.update({name: out_path})
|
||
logger.info('preprocess_handle success!')
|
||
|
||
# for name in para_names_geo:
|
||
# l1a_path = os.path.join(self.__workspace_preprocessed_path, name+".tif")
|
||
# self._tr.tran_geo_to_l1a(self.__preprocessed_paras[name], l1a_path, self.__preprocessed_paras['ori_sim'], is_class=False)
|
||
# self.__preprocessed_paras[name] = l1a_path
|
||
self.__cols_geo = self.imageHandler.get_img_width(self.__preprocessed_paras['sim_ori'])
|
||
self.__rows_geo = self.imageHandler.get_img_height(self.__preprocessed_paras['sim_ori'])
|
||
self.__cols = self.imageHandler.get_img_width(self.__preprocessed_paras['HH'])
|
||
self.__rows = self.imageHandler.get_img_height(self.__preprocessed_paras['HH'])
|
||
|
||
def __decompose(self, t3_path, rows, cols):
|
||
"""
|
||
极化分解：Freeman、Touzi、Yamaguchi、Cloude
|
||
:param t3_path: t3文件路径
|
||
:param rows: 影像行数
|
||
:return cols:影像列数
|
||
"""
|
||
# 计算特征组合
|
||
self.__getInputFeatures()
|
||
exeDir = ''
|
||
outFolderDic = {}
|
||
|
||
if 'Freeman' in self.__FeatureInput:
|
||
# freeman分解
|
||
freemanOutDir = os.path.join(self.__workspace_processing_path, 'freeman\\') # self.__workspace_processing_path + 'freeman\\'
|
||
freemDecom = PspFreemanDecomposition(
|
||
exeDir, t3_path, freemanOutDir)
|
||
flag = freemDecom.api_freeman_decomposition_T3(
|
||
0, 0, rows, cols)
|
||
if not flag:
|
||
logger.error('FreemanDecomposition err')
|
||
return False, None
|
||
outFolderDic['Freeman'] = freemanOutDir
|
||
|
||
if 'Touzi' in self.__FeatureInput:
|
||
# Touzi分解
|
||
touziOutDir = os.path.join(self.__workspace_processing_path, 'touzi\\') # self.__workspace_processing_path + 'touzi\\'
|
||
if not os.path.exists(touziOutDir):
|
||
os.makedirs(touziOutDir)
|
||
# touzi分解耗时较长，且对特征表达效果较差
|
||
p = PspTouziDecomposition(self.__preprocessed_paras, touziOutDir)
|
||
p.Touzi_decomposition_multiprocessing()
|
||
outFolderDic['Touzi'] = touziOutDir
|
||
|
||
if 'Yamaguchi' in self.__FeatureInput:
|
||
# Yamaguchi分解
|
||
yamaguchiOutDir = os.path.join(self.__workspace_processing_path, 'yamaguchi\\') # self.__workspace_processing_path + 'yamaguchi\\'
|
||
yamaguchiDecom = PspYamaguchiDecomposition(
|
||
exeDir, t3_path, yamaguchiOutDir)
|
||
flag = yamaguchiDecom.api_yamaguchi_4components_decomposition_T3(
|
||
0, 0, rows, cols)
|
||
if not flag:
|
||
logger.error('CloudePottierDecomposition err')
|
||
return False, None
|
||
outFolderDic['Yamaguchi'] = yamaguchiOutDir
|
||
|
||
if 'Cloude' in self.__FeatureInput:
|
||
# CloudePottier分解
|
||
cloudeOutDir = os.path.join(self.__workspace_processing_path, 'cloude\\') # self.__workspace_processing_path + 'cloude\\'
|
||
cloudeDecom = PspCloudePottierDecomposition(
|
||
exeDir, t3_path, cloudeOutDir)
|
||
flag = cloudeDecom.api_h_a_alpha_decomposition_T3(
|
||
0, 0, rows, cols)
|
||
if not flag:
|
||
logger.error('CloudePottierDecomposition err')
|
||
return False, None
|
||
outFolderDic['Cloude'] = cloudeOutDir
|
||
return True, outFolderDic
|
||
|
||
def __getInputFeatures(self):
|
||
"""
|
||
获取输入特征组合
|
||
1.Freeman：表面散射p_s、偶次散射p_d、体散射p_v 对应序号：[0,1,2]
|
||
2.Touzi：散射角α_s、散射相位ϕ_α、目标散射对称度τ、相对能量λ_i 对应序号：[3,4,5,6]
|
||
3.Yamaguchi：表面散射f_s、二次散射f_d、体散射f_v、螺旋体散射f_h 对应序号：[7,8,9,10]
|
||
4.Cloude - Pottier：分解散射熵H、反熵A、平均散射角α 对应序号：[11,12,13]
|
||
"""
|
||
decomposeList = self.__parameters_dic["FeatureCombination"].split(',')
|
||
stepSet = set()
|
||
for item in decomposeList:
|
||
stepSet.add(int(item))
|
||
|
||
for i in stepSet:
|
||
if 0 <= i < 3:
|
||
if 'Freeman' in self.__FeatureInput:
|
||
self.__FeatureInput['Freeman'].append(i)
|
||
else:
|
||
self.__FeatureInput['Freeman'] = [i]
|
||
elif 3 <= i < 7:
|
||
if 'Touzi' in self.__FeatureInput:
|
||
self.__FeatureInput['Touzi'].append(i)
|
||
else:
|
||
self.__FeatureInput['Touzi'] = [i]
|
||
elif 7 <= i < 11:
|
||
if 'Yamaguchi' in self.__FeatureInput:
|
||
self.__FeatureInput['Yamaguchi'].append(i)
|
||
else:
|
||
self.__FeatureInput['Yamaguchi'] = [i]
|
||
elif 11 <= i < 14:
|
||
if 'Cloude' in self.__FeatureInput:
|
||
self.__FeatureInput['Cloude'].append(i)
|
||
else:
|
||
self.__FeatureInput['Cloude'] = [i]
|
||
else:
|
||
logger.warning("__decomposeSteps Invalid data:%s!", i)
|
||
|
||
def __FeatureParaInit(self):
|
||
self.__FeatureMap["Freeman"] = [0, 1, 2]
|
||
self.__FeatureMap["Touzi"] = [3, 4, 5, 6]
|
||
self.__FeatureMap["Yamaguchi"] = [7, 8, 9, 10]
|
||
self.__FeatureMap["Cloude"] = [11, 12, 13]
|
||
|
||
self.___FeatureFileNameMap[0] = ['Freeman', "Freeman_Odd.bin"]
|
||
self.___FeatureFileNameMap[1] = ['Freeman', "Freeman_Dbl.bin"]
|
||
self.___FeatureFileNameMap[2] = ['Freeman', "Freeman_Vol.bin"]
|
||
self.___FeatureFileNameMap[3] = ['Touzi', "alpha.bin"]
|
||
self.___FeatureFileNameMap[4] = ['Touzi', "phi.bin"]
|
||
self.___FeatureFileNameMap[5] = ['Touzi', "tau.bin"]
|
||
self.___FeatureFileNameMap[6] = ['Touzi', "psi.bin"]
|
||
self.___FeatureFileNameMap[7] = ['Yamaguchi', "Yamaguchi4_Odd.bin"]
|
||
self.___FeatureFileNameMap[8] = ['Yamaguchi', "Yamaguchi4_Dbl.bin"]
|
||
self.___FeatureFileNameMap[9] = ['Yamaguchi', "Yamaguchi4_Vol.bin"]
|
||
self.___FeatureFileNameMap[10] = ['Yamaguchi', "Yamaguchi4_Hlx.bin"]
|
||
self.___FeatureFileNameMap[11] = ['Cloude', "anisotropy.bin"]
|
||
self.___FeatureFileNameMap[12] = ['Cloude', "entropy.bin"]
|
||
self.___FeatureFileNameMap[13] = ['Cloude', "alpha.bin"]
|
||
|
||
def create_roi(self, img_path):
|
||
"""
|
||
计算ROI掩膜
|
||
:return:掩膜路径
|
||
"""
|
||
processing_path = self.__workspace_processing_path
|
||
|
||
# 利用影像的有效范围生成MASK
|
||
tif_mask_path = os.path.join(processing_path, "tif_mask.tif") # processing_path + "tif_mask.tif"
|
||
alg.trans_tif2mask(tif_mask_path, img_path, np.nan)
|
||
logger.info('create ROI image success!')
|
||
return tif_mask_path
|
||
|
||
def gene_test_set(self, feature_tif_dic, optimal_feature):
|
||
"""
|
||
生成测试集
|
||
:param feature_tif_dic : 特征影像路径字典
|
||
:param optimal_feature : 最优特征子集
|
||
:return X_test_list : 分块测试集影像路径
|
||
"""
|
||
# 特征分块
|
||
bp = BlockProcess()
|
||
block_size = bp.get_block_size(self.__rows, self.__cols)
|
||
self.__block_size = block_size
|
||
|
||
bp.cut_new(feature_tif_dic, self.__workspace_block_tif_path, ['tif', 'tiff'], 'tif', block_size)
|
||
img_dir, img_name = bp.get_file_names(self.__workspace_block_tif_path, ['tif'])
|
||
dir_dict_all = bp.get_same_img(img_dir, img_name)
|
||
|
||
# 选择最优特征子集特征影像
|
||
dir_dict = {}
|
||
for n, key in zip(range(len(dir_dict_all)),dir_dict_all):
|
||
if n in optimal_feature:
|
||
dir_dict.update({key: dir_dict_all[key]})
|
||
logger.info('test_feature:%s', dir_dict.keys())
|
||
logger.info('blocking tifs success!')
|
||
X_test_list = []
|
||
# 特征维度合并
|
||
for key in dir_dict:
|
||
key_name = key
|
||
block_num = len(dir_dict[key])
|
||
break
|
||
for n in range(block_num):
|
||
name = os.path.basename(dir_dict[key_name][n])
|
||
suffix = '_' + name.split('_')[-4] + "_" + name.split(
|
||
'_')[-3] + "_" + name.split('_')[-2] + "_" + name.split('_')[-1]
|
||
features_path = os.path.join(self.__workspace_block_tif_processed_path, "features","features" + suffix) # self.__workspace_block_tif_processed_path + "features\\features" + suffix
|
||
X_test_list.append(features_path)
|
||
features_array = np.zeros(
|
||
(len(dir_dict), block_size, block_size), dtype='float32')
|
||
for m, value in zip(range(len(dir_dict)), dir_dict.values()):
|
||
features_array[m, :, :] = ImageHandler.get_band_array(
|
||
value[n])
|
||
features_array[np.isnan(features_array)] = 0.0 # 异常值转为0
|
||
ImageHandler.write_img(features_path, '', [0, 0, 0, 0, 0, 0], features_array)
|
||
logger.info('create features matrix success!')
|
||
return X_test_list
|
||
|
||
def predict_blok(self, clf, X_test, rows, cols, img_path, row_begin, col_begin, block_sum, n):
|
||
Y_test = clf.predict(X_test)
|
||
img = Y_test.reshape(rows, cols)
|
||
out_image = Image.fromarray(img)
|
||
out_image.save(img_path)
|
||
# bp = BlockProcess()
|
||
# bp.assign_spatial_reference_bypoint(row_begin, col_begin, self.__proj, self.__geo, img_path)
|
||
sr = osr.SpatialReference()
|
||
sr.ImportFromWkt(self.__proj)
|
||
geo_transform = (self.__geo[0] + col_begin * self.__geo[1] + row_begin * self.__geo[2],
|
||
self.__geo[1],
|
||
self.__geo[2],
|
||
self.__geo[3] + col_begin * self.__geo[4] + row_begin * self.__geo[5],
|
||
self.__geo[4],
|
||
self.__geo[5]
|
||
)
|
||
dst_ds = gdal.Open(img_path, gdal.GA_Update)
|
||
if dst_ds is None:
|
||
return False
|
||
dst_ds.SetProjection(sr.ExportToWkt())
|
||
dst_ds.SetGeoTransform(geo_transform)
|
||
del dst_ds
|
||
logger.info('total:%s,block:%s test data finished !', block_sum, n)
|
||
return True
|
||
|
||
def predict(self, clf, X_test_list):
|
||
"""
|
||
预测数据
|
||
:param clf : svm模型
|
||
:return X_test_list: 分块测试集影像路径
|
||
"""
|
||
# 开启多进程处理
|
||
bp = BlockProcess()
|
||
out_tif_name = 'landcover'
|
||
block_features_dir = X_test_list
|
||
bp_cover_dir = os.path.join(self.__workspace_block_tif_processed_path, out_tif_name + '\\') # self.__workspace_block_tif_processed_path + out_tif_name + '\\'
|
||
if not os.path.exists(bp_cover_dir):
|
||
os.makedirs(bp_cover_dir)
|
||
|
||
processes_num = min([len(block_features_dir), 10])
|
||
pool = multiprocessing.Pool(processes=processes_num)
|
||
pl = []
|
||
for path, n in zip(block_features_dir, range(len(block_features_dir))):
|
||
name = os.path.split(path)[1]
|
||
features_array = self.imageHandler.get_data(path)
|
||
|
||
X_test = np.reshape(features_array, (features_array.shape[0], features_array[0].size)).T
|
||
|
||
suffix = '_' + name.split('_')[-4] + "_" + name.split('_')[-3] + "_" + name.split('_')[-2] + "_" + name.split('_')[-1]
|
||
img_path = os.path.join(bp_cover_dir, out_tif_name + suffix) # bp_cover_dir + out_tif_name + suffix
|
||
row_begin = int(name.split('_')[-4])
|
||
col_begin = int(name.split('_')[-2])
|
||
pl.append(pool.apply_async(self.predict_blok, (clf, X_test, self.__block_size, self.__block_size, img_path, row_begin, col_begin,len(block_features_dir), n)))
|
||
logger.info('total:%s,block:%s testing data !', len(block_features_dir), n)
|
||
pool.close()
|
||
pool.join()
|
||
|
||
# 合并影像
|
||
data_dir = bp_cover_dir
|
||
out_path = self.__workspace_processing_path[0:-1]
|
||
bp.combine_new(
|
||
data_dir,
|
||
self.__cols,
|
||
self.__rows,
|
||
out_path,
|
||
file_type=['tif'],
|
||
datetype='float32')
|
||
|
||
# 添加地理信息
|
||
cover_path = os.path.join(self.__workspace_processing_path, out_tif_name + ".tif") # self.__workspace_processing_path + out_tif_name + ".tif"
|
||
bp.assign_spatial_reference_byfile(self.__ref_img_path, cover_path)
|
||
return cover_path
|
||
|
||
def ahv_to_t3(self):
|
||
# 全极化tif转bin格式T3数据
|
||
atp = AHVToPolsarpro()
|
||
t3_path = os.path.join(self.__workspace_processing_path, 'psp_t3\\') # self.__workspace_processing_path + 'psp_t3\\'
|
||
# atp.ahv_to_polsarpro_t3(t3_path, self.__workspace_preprocessed_path)
|
||
|
||
polarization = ['HH', 'HV', 'VH', 'VV']
|
||
calibration = Calibration.get_Calibration_coefficient(self.__processing_paras['Origin_META'], polarization)
|
||
tif_path = atp.calibration(calibration, in_ahv_dir=self.__workspace_preprocessed_path)
|
||
atp.ahv_to_polsarpro_t3_soil(t3_path, tif_path)
|
||
|
||
# Lee滤波
|
||
leeFilter = LeeRefinedFilterT3()
|
||
lee_filter_path = os.path.join(self.__workspace_processing_path, 'lee_filter\\') # self.__workspace_processing_path + 'lee_filter\\'
|
||
|
||
leeFilter.api_lee_refined_filter_T3('', t3_path, lee_filter_path, 0, 0, atp.rows(), atp.cols(), FILTER_SIZE)
|
||
logger.info("refine_lee filter success!")
|
||
logger.info('progress bar: 30%')
|
||
return lee_filter_path
|
||
|
||
def calInterpolation_bil_Wgs84_rc_sar_sigma(self, parameter_path, dem_rc, in_sar, out_sar):
|
||
'''
|
||
# std::cout << "mode 11";
|
||
# std::cout << "SIMOrthoProgram.exe 11 in_parameter_path in_rc_wgs84_path in_ori_sar_path out_orth_sar_path";
|
||
'''
|
||
exe_path = r".\baseTool\x64\Release\SIMOrthoProgram-L-SAR.exe"
|
||
exe_cmd = r"set PROJ_LIB=.\baseTool\x64\Release; & {0} {1} {2} {3} {4} {5}".format(exe_path, 11, parameter_path,
|
||
dem_rc, in_sar, out_sar)
|
||
print(exe_cmd)
|
||
print(os.system(exe_cmd))
|
||
print("==========================================================================")
|
||
|
||
def features_geo(self, features_path):
|
||
dir = os.path.join(self.__workspace_processing_path, 'features_geo\\')
|
||
if not os.path.exists(dir):
|
||
os.mkdir(dir)
|
||
for key, file in zip(features_path, features_path.values()):
|
||
name = key + '_geo.tif'
|
||
out_path = os.path.join(dir, name)
|
||
self.calInterpolation_bil_Wgs84_rc_sar_sigma(self.__processing_paras['paraMeter'],
|
||
self.__preprocessed_paras['sim_ori'], file, out_path)
|
||
return dir
|
||
|
||
def process_handle(self, start):
|
||
"""
|
||
算法主处理函数
|
||
"""
|
||
hh_geo_path = self.__workspace_processing_path + "hh_geo.tif"
|
||
self.calInterpolation_bil_Wgs84_rc_sar_sigma(self.__processing_paras['paraMeter'],
|
||
self.__preprocessed_paras['sim_ori'],
|
||
self.__preprocessed_paras['HH'], hh_geo_path)
|
||
|
||
# 读取实测值，获取多边形区域内所有的点，分为训练集数据和测试集数据
|
||
pm = LandCoverMeasCsv(self.__processing_paras['LabelData'], hh_geo_path, MAX_TRAN_NUM)
|
||
train_data_list = pm.api_read_measure()
|
||
train_data_dic = csvh.trans_landCover_list2dic(train_data_list)
|
||
|
||
csvh_roi = csvHandle(self.__rows_geo, self.__cols_geo)
|
||
# train_data_dic = csvh_roi.trans_landCover_measuredata_dic(csvh_roi.readcsv(self.__processing_paras['LabelData']), self.__preprocessed_paras['ori_sim'], MAX_TRAN_NUM)
|
||
label_img = csvh_roi.get_roi_img()
|
||
if(len(label_img) != 0):
|
||
self.imageHandler.write_img(os.path.join(self.__workspace_processing_path, "label_img.tif"),"",[0,0,0,0,0,0],label_img)
|
||
|
||
logger.info("read csv data success!")
|
||
logger.info('progress bar: 20%')
|
||
# 极化分解
|
||
flag, outFolder = self.__decompose(self.ahv_to_t3(), self.__rows, self.__cols)
|
||
|
||
feature_tif_paths = {}
|
||
for key in outFolder:
|
||
feature_bin_dic = outFolder[key]
|
||
if key == 'Touzi':
|
||
for path in list(glob.glob(os.path.join(feature_bin_dic, '*.tif'))):
|
||
name = os.path.split(path)[1].split('.')[0]
|
||
shutil.copyfile(path, os.path.join(self.__feature_tif_dir, name +'.tif') ) # self.__feature_tif_dir +'\\'+name+'.tif')
|
||
feature_tif_paths.update({name: os.path.join(self.__feature_tif_dir, name+'.tif')}) # self.__feature_tif_dir +'\\'+name+'.tif'})
|
||
else:
|
||
feature_tif_paths.update(write_bin_to_tif(self.__feature_tif_dir, feature_bin_dic))
|
||
|
||
logging.info("feature_tif_paths:%s",feature_tif_paths)
|
||
# 对所有特征进行地理编码
|
||
feature_geo = self.features_geo(feature_tif_paths)
|
||
# 新添加的特征做归一化
|
||
# for name in self.__feature_name_list:
|
||
# proj, geo, arr = self.imageHandler.read_img(self.__preprocessed_paras[name])
|
||
# arr = ml.standardization(arr)
|
||
# self.imageHandler.write_img(os.path.join(self.__feature_tif_dir, name+".tif"), proj, geo, arr)
|
||
|
||
logger.info("decompose feature success!")
|
||
logger.info('progress bar: 50%')
|
||
|
||
# 生成最优特征子集训练集
|
||
X_train, Y_train, optimal_feature = ml.gene_optimal_train_set(train_data_dic, feature_geo, 0.07, 0.85)
|
||
|
||
# 训练模型
|
||
# cost = self.__processing_paras["Cost"]
|
||
# kernel = self.__processing_paras["Kernel"]
|
||
# if cost < 0:
|
||
# cost = 1
|
||
# if kernel != 'rbf' and kernel != 'poly' and kernel != 'linear' and kernel != 'sigmoid':
|
||
# kernel = 'rbf'
|
||
|
||
# RF
|
||
clf = ml.trainRF(X_train, Y_train)
|
||
logger.info("RF train successful")
|
||
|
||
# # SVM
|
||
# clf = ml.trainSVM(X_train, Y_train, cost, kernel)
|
||
# logger.info("SVM train successful")
|
||
logger.info('progress bar: 60%')
|
||
|
||
# 生成测试集
|
||
# X_test_path_list = ml.gene_test_set(self.__feature_tif_dir, optimal_feature)
|
||
X_test_path_list = ml.gene_test_set(feature_geo, optimal_feature)
|
||
# 预测
|
||
logger.info('testing')
|
||
cover_path = ml.predict(clf, X_test_path_list, EXE_NAME, self.__workspace_processing_path, self.__rows_geo, self.__cols_geo)
|
||
logger.info('test success!')
|
||
logger.info('progress bar: 95%')
|
||
|
||
# 转换数据
|
||
proj, geo, cover_data = self.imageHandler.read_img(cover_path)
|
||
|
||
# 形态学（闭运算）去roi区域噪点
|
||
cover_data = np.uint8(cover_data)
|
||
kernel = np.ones((5, 5), np.uint8)
|
||
cover_data = cv2.erode(cv2.dilate(cover_data, kernel), kernel)
|
||
cover_data = np.int16(cover_data)
|
||
for id, class_id in zip(train_data_dic['ids'], train_data_dic['class_ids']):
|
||
cover_data[np.where(cover_data == id)] = class_id
|
||
|
||
# cover_path = cover_path.replace('.tif', '_geo.tif')
|
||
cover_geo_path = cover_path.replace('.tif', '_geo.tif')
|
||
self.imageHandler.write_img(cover_geo_path, proj, geo, cover_data)
|
||
|
||
|
||
# l1a图像坐标转换地理坐标
|
||
# self.calInterpolation_bil_Wgs84_rc_sar_sigma(self.__processing_paras['paraMeter'],
|
||
# self.__preprocessed_paras['sim_ori'], cover_path, cover_geo_path)
|
||
# # self._tr.l1a_2_geo_int(self.__preprocessed_paras['ori_sim'], cover_path, cover_geo_path, 'nearest')
|
||
proj, geo, cover_data_geo = self.imageHandler.read_img(cover_geo_path)
|
||
|
||
|
||
proj, geo, cover_data_hh = self.imageHandler.read_img(hh_geo_path)
|
||
# self._tr.l1a_2_geo_int(self.__preprocessed_paras['ori_sim'], self.__preprocessed_paras['HH'], hh_geo_path)
|
||
roi_img = self.imageHandler.get_band_array(self.create_roi(hh_geo_path))
|
||
|
||
# 获取影像roi区域
|
||
cover_data_geo = cover_data_geo * roi_img
|
||
product_dir = os.path.join(self.__workspace_processing_path, 'product\\') # self.__workspace_processing_path + 'product\\'
|
||
SrcImageName = os.path.split(self.__input_paras["AHV"]['ParaValue'])[1].split('.tar.gz')[0] + tar + '.tif'
|
||
product_path = os.path.join(product_dir, SrcImageName) # product_dir + 'LandCoverProduct.tif'
|
||
self.imageHandler.write_img(product_path, proj, geo, np.int16(cover_data_geo), no_data=0)
|
||
|
||
# 生成快视图
|
||
self.imageHandler.write_quick_view(product_path, color_img=True)
|
||
|
||
self.create_meta_file(product_path)
|
||
# 文件夹打包
|
||
file.make_targz(self.__out_para, product_dir)
|
||
logger.info('process_handle success!')
|
||
logger.info('successful production of ' + EXE_NAME + ' products!')
|
||
|
||
def create_meta_file(self, product_path):
|
||
# 生成产品元文件
|
||
product_dir = os.path.join(self.__workspace_processing_path, 'product\\') # self.__workspace_processing_path + 'product\\'
|
||
# xml_path = "./model_meta.xml"
|
||
tem_folder = os.path.join(self.__workspace_path, EXE_NAME, "Temporary\\") # self.__workspace_path + EXE_NAME + r"\Temporary""\\"
|
||
image_path = product_path
|
||
out_path1 = os.path.join(tem_folder, "trans_geo_projcs.tif")
|
||
out_path2 = os.path.join(tem_folder, "trans_projcs_geo.tif")
|
||
SrcImageName = os.path.split(self.__input_paras["AHV"]['ParaValue'])[1].split('.tar.gz')[0]
|
||
|
||
# par_dict = CreateDict(image_path, self.processinfo, out_path1, out_path2).calu_nature(start)
|
||
# model_xml_path = os.path.join(tem_folder, "creat_standard.meta.xml") # 输出xml路径
|
||
# CreateStadardXmlFile(xml_path, self.alg_xml_path, par_dict, model_xml_path).create_standard_xml()
|
||
#
|
||
# meta_xml_path = os.path.join(product_dir, EXE_NAME + "Product.meta.xml") # product_dir+EXE_NAME+"Product.meta.xml"
|
||
# type_id_name, type_id_parent = csvh.class_landcover_list(self.__processing_paras['LabelData'])
|
||
# CreateMetafile(self.__processing_paras['META'], self.alg_xml_path, model_xml_path, meta_xml_path).process2(type_id_name, type_id_parent, SrcImageName)
|
||
|
||
model_path = "./product.xml"
|
||
meta_xml_path = os.path.join(product_dir, SrcImageName + tar + ".meta.xml")
|
||
|
||
para_dict = CreateMetaDict(image_path, self.__processing_paras['Origin_META'], product_dir,
|
||
out_path1, out_path2).calu_nature()
|
||
para_dict.update({"imageinfo_ProductName": "地表覆盖类型"})
|
||
para_dict.update({"imageinfo_ProductIdentifier": "LandCover"})
|
||
para_dict.update({"imageinfo_ProductLevel": productLevel})
|
||
para_dict.update({"ProductProductionInfo_BandSelection": "1,2"})
|
||
para_dict.update({"ProductProductionInfo_AuxiliaryDataDescription": "LabelData"})
|
||
CreateProductXml(para_dict, model_path, meta_xml_path).create_standard_xml()
|
||
|
||
temp_folder = os.path.join(self.__workspace_path, EXE_NAME, 'Output')
|
||
out_xml = os.path.join(temp_folder, os.path.basename(meta_xml_path))
|
||
if os.path.exists(temp_folder) is False:
|
||
os.mkdir(temp_folder)
|
||
# CreateProductXml(para_dict, model_path, out_xml).create_standard_xml()
|
||
shutil.copy(meta_xml_path, out_xml)
|
||
|
||
if __name__ == '__main__':
|
||
multiprocessing.freeze_support() #解决打包与运行错误
|
||
start = datetime.datetime.now()
|
||
try:
|
||
if len(sys.argv) < 2:
|
||
xml_path = 'LandCover-L-SAR.xml'
|
||
else:
|
||
xml_path = sys.argv[1]
|
||
main_handler = LandCoverMain(xml_path)
|
||
if main_handler.check_source() is False:
|
||
raise Exception('check_source() failed!')
|
||
if main_handler.preprocess_handle() is False:
|
||
raise Exception('preprocess_handle() failed!')
|
||
if main_handler.process_handle(start) is False:
|
||
raise Exception('process_handle() failed!')
|
||
except Exception:
|
||
logger.exception("run-time error!")
|
||
finally:
|
||
main_handler.del_temp_workspace()
|
||
|
||
end = datetime.datetime.now()
|
||
msg = 'running use time: %s ' % (end - start)
|
||
logger.info(msg)
|
||
|
||
# 模型的导入导出 https://blog.csdn.net/daxiaofan/article/details/71189015?ops_request_misc=&request_id=&biz_id=102&utm_term=scikit-learn%20%E6%A8%A1%E5%9E%8B%E5%AF%BC%E5%87%BA%E5%AF%BC%E5%85%A5&utm_medium=distribute.pc_search_result.none-task-blog-2~all~sobaiduweb~default-5-71189015.142^v59^js_top,201^v3^control_1&spm=1018.2226.3001.4187 |