修改分块大小，增加效率，正射新增裁剪dem，防止输入dem范围过大，影响效率
tian jiax
parent
90e9d6678c
commit
dcc953152b
|
|
@ -1,7 +1,7 @@
|
|||
<?xml version='1.0' encoding='utf-8'?>
|
||||
<Root>
|
||||
<TaskID>CSAR_202107275419_0001-0</TaskID>
|
||||
<WorkSpace>E:\Result_GF3\</WorkSpace>
|
||||
<WorkSpace>D:\micro\WorkSpace\</WorkSpace>
|
||||
<AlgCompt>
|
||||
<DataTransModel>File</DataTransModel>
|
||||
<Artificial>ElementAlg</Artificial>
|
||||
|
|
@ -45,7 +45,7 @@
|
|||
<ParaType>File</ParaType>
|
||||
<DataType>tar.gz</DataType>
|
||||
<ParaSource>Cal</ParaSource>
|
||||
<ParaValue>E:\GF3Data\vegetationPhenology\GF3C_MYC_QPSI_006270_E100.4_N27.0_20230615_L1A_AHV_L10000158764.tar.gz</ParaValue>
|
||||
<ParaValue>D:\BaiduNetdiskDownload\西北新增CSAR盐碱度\GF3C_KSC_QPSI_008440_E86.2_N44.1_20231113_L1A_AHV_L10000215823.tar.gz</ParaValue>
|
||||
<EnModification>True</EnModification>
|
||||
<EnMultipleChoice>False</EnMultipleChoice>
|
||||
<Control>File</Control>
|
||||
|
|
@ -58,9 +58,9 @@
|
|||
<ParaChsName>DEM数字高程影像</ParaChsName>
|
||||
<Description>30m分辨率DEM数字高程影像tif</Description>
|
||||
<ParaType>File</ParaType>
|
||||
<DataType>zip</DataType>
|
||||
<DataType>File</DataType>
|
||||
<ParaSource>Cal</ParaSource>
|
||||
<ParaValue>E:\GF3Data\vegetationPhenology\CSAR_vegetation_ASTGTM2_N26E100_dem.zip;E:\GF3Data\vegetationPhenology\CSAR_vegetation_ASTGTM2_N27E100_dem.zip</ParaValue>
|
||||
<ParaValue>D:\BaiduNetdiskDownload\dem_xb</ParaValue>
|
||||
<EnModification>True</EnModification>
|
||||
<EnMultipleChoice>True</EnMultipleChoice>
|
||||
<Control>File</Control>
|
||||
|
|
@ -92,7 +92,7 @@
|
|||
<ParaType>File</ParaType>
|
||||
<DataType>tar.gz</DataType>
|
||||
<ParaSource>Cal</ParaSource>
|
||||
<ParaValue>E:\Result_GF3\Ortho\Output\GF3C_MYC_QPSI_006270_E100.4_N27.0_20230615_L1A_AHV_L10000158764-ortho.tar.gz</ParaValue>
|
||||
<ParaValue>D:\micro\WorkSpace\Ortho\Output\GF3C_KSC_QPSI_008440_E86.2_N44.1_20231113_L1A_AHV_L10000215823-ortho.tar.gz</ParaValue>
|
||||
<MaxValue>DEFAULT</MaxValue>
|
||||
<MinValue>DEFAULT</MinValue>
|
||||
<OptionValue>DEFAULT</OptionValue>
|
||||
|
|
|
|||
|
|
@ -13,6 +13,7 @@ import logging
|
|||
from tool.algorithm.block.blockprocess import BlockProcess
|
||||
from tool.algorithm.image.ImageHandle import ImageHandler
|
||||
from tool.algorithm.xml.CreateMetaDict import CreateMetaDict, CreateProductXml, OrthoAzimuth
|
||||
from tool.algorithm.xml.AnalysisXml import DictXml
|
||||
from tool.algorithm.algtools.PreProcess import PreProcess as pp
|
||||
import tarfile
|
||||
from tool.algorithm.xml.AlgXmlHandle import ManageAlgXML, CheckSource # 导入xml文件读取与检查文件
|
||||
|
|
@ -498,14 +499,19 @@ class OrthoMain:
|
|||
logger.info('progress bar :100％')
|
||||
return True
|
||||
|
||||
|
||||
pass
|
||||
|
||||
|
||||
def cut_dem(self, dem_merged_path, meta_file_path):
|
||||
left_up_lon = 0
|
||||
left_up_lat = 0
|
||||
|
||||
_, scopes = DictXml(meta_file_path).get_extend()
|
||||
intersect_polygon = pp().intersect_polygon(scopes)
|
||||
if intersect_polygon is None:
|
||||
raise Exception('cal intersect box fail!')
|
||||
shp_path = os.path.join(self.__workspace_Temporary_path, 'IntersectPolygon.shp')
|
||||
if pp().write_polygon_shp(shp_path, intersect_polygon, 4326) is False:
|
||||
raise Exception('create intersect shp fail!')
|
||||
dem_process = os.path.join(self.__workspace_Temporary_path, 'dem_cut.tif')
|
||||
pp().cut_img(dem_process, dem_merged_path, shp_path)
|
||||
return dem_process
|
||||
|
||||
def process_sim_ori(self, ori_sim, sim_ori):
|
||||
p = pp()
|
||||
|
|
@ -534,7 +540,7 @@ class OrthoMain:
|
|||
out_dem_path = self.__workspace_ResampledDEM_path
|
||||
dem_merged_path=DEMProcess.dem_merged(in_dem_path, meta_file_path, out_dem_path) # 生成TestDEM\mergedDEM_VRT.tif
|
||||
|
||||
# self.cut_dem(dem_merged_path, meta_file_path)
|
||||
dem_path = self.cut_dem(dem_merged_path, meta_file_path)
|
||||
# 2、间接定位法求解行列坐标
|
||||
slc_paths = self.__in_processing_paras["SLC"]
|
||||
# 2.1 生成映射表
|
||||
|
|
@ -553,7 +559,7 @@ class OrthoMain:
|
|||
in_slc_path=os.path.join(slc_paths,slc_path)
|
||||
break
|
||||
# 获取校正模型后
|
||||
Orthorectification.preCaldem_sar_rc(dem_merged_path,in_slc_path,self.__workspace_Temporary_path,self.__workspace_package_path.replace("\\","\\\\")) # 初步筛选坐标范围
|
||||
Orthorectification.preCaldem_sar_rc(dem_path,in_slc_path,self.__workspace_Temporary_path,self.__workspace_package_path.replace("\\","\\\\")) # 初步筛选坐标范围
|
||||
|
||||
logger.info('progress bar: 40%')
|
||||
# clip_dem_reample_path=os.path.join(self.__workspace_Temporary_path, "SAR_dem.tiff")
|
||||
|
|
@ -686,12 +692,12 @@ class OrthoMain:
|
|||
# 生成压缩包
|
||||
logger.info('progress bar :94％')
|
||||
logger.info('start make targz..')
|
||||
self.del_floder(self.__workspace_unpack_path)
|
||||
self.del_floder(self.__workspace_ResampledDEM_path)
|
||||
self.del_floder(self.__workspace_LutImg_path)
|
||||
self.del_floder(self.__workspace_IncidenceImg_path)
|
||||
self.del_floder(self.__workspace_SimImg_path)
|
||||
self.del_floder(self.__workspace_SARIntensity_path)
|
||||
# self.del_floder(self.__workspace_unpack_path)
|
||||
# self.del_floder(self.__workspace_ResampledDEM_path)
|
||||
# self.del_floder(self.__workspace_LutImg_path)
|
||||
# self.del_floder(self.__workspace_IncidenceImg_path)
|
||||
# self.del_floder(self.__workspace_SimImg_path)
|
||||
# self.del_floder(self.__workspace_SARIntensity_path)
|
||||
self.make_targz(self.__out_para, self.__workspace_package_path+"\\")
|
||||
logger.info('make targz finish')
|
||||
logger.info('progress bar :100％')
|
||||
|
|
|
|||
|
|
@ -12,6 +12,7 @@ import logging
|
|||
from tool.algorithm.algtools.logHandler import LogHandler
|
||||
from tool.algorithm.block.blockprocess import BlockProcess
|
||||
from tool.algorithm.xml.AlgXmlHandle import ManageAlgXML, CheckSource
|
||||
from tool.algorithm.xml.AnalysisXml import DictXml
|
||||
from tool.algorithm.xml.CreateMetaDict import CreateMetaDict, CreateProductXml, OrthoAzimuth
|
||||
from tool.algorithm.image.ImageHandle import ImageHandler
|
||||
from tool.algorithm.algtools.PreProcess import PreProcess as pp
|
||||
|
|
@ -277,6 +278,18 @@ class ScatteringMain:
|
|||
pp().cut_img(sim_ori_process, sim_ori, shp_path)
|
||||
return sim_ori_process
|
||||
|
||||
def cut_dem(self, dem_merged_path, meta_file_path):
|
||||
_, scopes = DictXml(meta_file_path).get_extend()
|
||||
intersect_polygon = pp().intersect_polygon(scopes)
|
||||
if intersect_polygon is None:
|
||||
raise Exception('cal intersect box fail!')
|
||||
shp_path = os.path.join(self.__workspace_preprocessing_path, 'IntersectPolygon.shp')
|
||||
if pp().write_polygon_shp(shp_path, intersect_polygon, 4326) is False:
|
||||
raise Exception('create intersect shp fail!')
|
||||
dem_process = os.path.join(self.__workspace_preprocessing_path, 'dem_cut.tif')
|
||||
pp().cut_img(dem_process, dem_merged_path, shp_path)
|
||||
return dem_process
|
||||
|
||||
def process_handle(self, start):
|
||||
in_tif_paths = list(glob.glob(os.path.join(self.__in_processing_paras['SLC'], '*.tif')))
|
||||
if in_tif_paths == []:
|
||||
|
|
@ -319,6 +332,7 @@ class ScatteringMain:
|
|||
|
||||
dem_merged_path = DEMProcess.dem_merged(in_dem_path, meta_file_path,
|
||||
out_dem_path) # 生成TestDEM\mergedDEM_VRT.tif
|
||||
dem_path = self.cut_dem(dem_merged_path, meta_file_path)
|
||||
|
||||
in_slc_path = None
|
||||
for slc_path in in_tif_paths:
|
||||
|
|
@ -328,7 +342,7 @@ class ScatteringMain:
|
|||
break
|
||||
|
||||
# 获取校正模型后
|
||||
Orthorectification.preCaldem_sar_rc(dem_merged_path, in_slc_path, self.__workspace_preprocessing_path,
|
||||
Orthorectification.preCaldem_sar_rc(dem_path, in_slc_path, self.__workspace_preprocessing_path,
|
||||
self.__workspace_processing_path.replace("\\", "\\\\")) # 初步筛选坐标范围
|
||||
logger.info('progress bar: 40%')
|
||||
# clip_dem_reample_path=os.path.join(self.__workspace_preprocessing_path, "SAR_dem.tiff")
|
||||
|
|
|
|||
|
|
@ -39,7 +39,7 @@
|
|||
<DataType>tar.gz</DataType>
|
||||
<ParaSource>Man</ParaSource>
|
||||
<ParaValue>
|
||||
F:\2023xibei\GF3C_KSC_QPSI_008440_E86.0_N44.7_20231113_L1A_AHV_L10000215825-ortho.tar.gz
|
||||
F:\2024xibei\GF3B_KSC_QPSI_010328_E86.1_N44.3_20231109_L1A_AHV_L10000262133-ortho.tar.gz
|
||||
</ParaValue>
|
||||
<MaxValue>DEFAULT</MaxValue>
|
||||
<MinValue>DEFAULT</MinValue>
|
||||
|
|
@ -110,7 +110,7 @@
|
|||
<ParaType>File</ParaType>
|
||||
<DataType>tif</DataType>
|
||||
<ParaSource>Man</ParaSource>
|
||||
<ParaValue>F:\2023xibei\S2_NDVImed2.tif</ParaValue>
|
||||
<ParaValue>F:\2024xibei\S2_NDVImed.tif</ParaValue>
|
||||
<MaxValue>DEFAULT</MaxValue>
|
||||
<MinValue>DEFAULT</MinValue>
|
||||
<OptionValue>DEFAULT</OptionValue>
|
||||
|
|
@ -140,7 +140,7 @@
|
|||
<ParaType>File</ParaType>
|
||||
<DataType>tar.gz</DataType>
|
||||
<ParaSource>Man</ParaSource>
|
||||
<ParaValue>D:\micro\WorkSpace\SoilSalinityPredict\Output\GF3C_KSC_QPSI_008440_E86.0_N44.7_20231113_L1A_AHV_L10000215825-ortho-SSAA.tar.gz</ParaValue>
|
||||
<ParaValue>D:\micro\WorkSpace\SoilSalinityPredict\Output\GF3B_KSC_QPSI_010328_E86.1_N44.3_20231109_L1A_AHV_L10000262133-ortho-SSAA.tar.gz</ParaValue>
|
||||
<MaxValue>DEFAULT</MaxValue>
|
||||
<MinValue>DEFAULT</MinValue>
|
||||
<OptionValue>DEFAULT</OptionValue>
|
||||
|
|
|
|||
|
|
@ -1,7 +1,7 @@
|
|||
<?xml version='1.0' encoding='utf-8'?>
|
||||
<Root>
|
||||
<TaskID>CSAR_202107275419_0001-0</TaskID>
|
||||
<WorkSpace>D:\soilSanlinity\</WorkSpace>
|
||||
<WorkSpace>D:\micro\WorkSpace\</WorkSpace>
|
||||
<AlgCompt>
|
||||
<DataTransModel>File</DataTransModel>
|
||||
<Artificial>ElementAlg</Artificial>
|
||||
|
|
@ -80,7 +80,7 @@
|
|||
<ParaType>File</ParaType>
|
||||
<DataType>tif</DataType>
|
||||
<ParaSource>Man</ParaSource>
|
||||
<ParaValue>E:\C-SARNDVINDWIfugaileixingDEM\fugaileixing.tif</ParaValue>
|
||||
<ParaValue>F:\2023xibei\fugaileixing.tif</ParaValue>
|
||||
<MaxValue>DEFAULT</MaxValue>
|
||||
<MinValue>DEFAULT</MinValue>
|
||||
<OptionValue>DEFAULT</OptionValue>
|
||||
|
|
@ -108,7 +108,7 @@
|
|||
<ParaType>File</ParaType>
|
||||
<DataType>tif</DataType>
|
||||
<ParaSource>Man</ParaSource>
|
||||
<ParaValue>E:\C-SARNDVINDWIfugaileixingDEM\S2_NDVImed2.tif</ParaValue>
|
||||
<ParaValue>F:\2023xibei\S2_NDVImed2.tif</ParaValue>
|
||||
<MaxValue>DEFAULT</MaxValue>
|
||||
<MinValue>DEFAULT</MinValue>
|
||||
<OptionValue>DEFAULT</OptionValue>
|
||||
|
|
@ -138,7 +138,7 @@
|
|||
<ParaType>File</ParaType>
|
||||
<DataType>tar.gz</DataType>
|
||||
<ParaSource>Man</ParaSource>
|
||||
<ParaValue>D:\soilSanlinity\SoilSalinity\Output\GF3C_KSC_QPSI_008440_E86.0_N44.7_20231113_L1A_AHV_L10000215825-ortho-SSAA.tar.gz</ParaValue>
|
||||
<ParaValue>D:\micro\WorkSpace\SoilSalinity\Output\GF3C_KSC_QPSI_008440_E86.0_N44.7_20231113_L1A_AHV_L10000215825-ortho-SSAA.tar.gz</ParaValue>
|
||||
<MaxValue>DEFAULT</MaxValue>
|
||||
<MinValue>DEFAULT</MinValue>
|
||||
<OptionValue>DEFAULT</OptionValue>
|
||||
|
|
|
|||
|
|
@ -315,6 +315,7 @@ class SalinityMain:
|
|||
rows = self.imageHandler.get_img_height(self.__preprocessed_paras['HH'])
|
||||
cols = self.imageHandler.get_img_width(self.__preprocessed_paras['HH'])
|
||||
block_size = bp.get_block_size(rows, cols)
|
||||
logger.info('block size is :%s', block_size)
|
||||
bp.cut(out_dir, 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 = bp.get_same_img(img_dir, img_name)
|
||||
|
|
@ -337,6 +338,7 @@ class SalinityMain:
|
|||
features_array[np.isnan(features_array)] = 0.0
|
||||
features_array[np.isinf(features_array)] = 0.0
|
||||
self.imageHandler.write_img(features_path, "", [0, 0, 1, 0, 0, 1], features_array)
|
||||
|
||||
logger.info('create features matrix success!')
|
||||
|
||||
# for n in range(block_num):
|
||||
|
|
|
|||
|
|
@ -8,7 +8,6 @@ productLevel = 5
|
|||
exe_name = SoilSalinity
|
||||
# 开启调试模式则不删除临时工作区，True:开启调试，False:不开启调试
|
||||
debug = False
|
||||
|
||||
######2-算法处理参数######
|
||||
pixelspace = 5
|
||||
######3-算法结果参数######
|
||||
|
|
|
|||
|
|
@ -98,6 +98,7 @@ class PspHAAlphaDecomposition:
|
|||
self.api_trans_T3_to_tif(h_a_alpha_out_dir, polsarpro_in_dir)
|
||||
if is_read_to_dic:
|
||||
h_a_alpha_features.update(self.api_read_T3_matrix(polsarpro_in_dir))
|
||||
logger.info('run bin to tif success!')
|
||||
return h_a_alpha_features
|
||||
|
||||
def api_h_a_alpha_decomposition_T3(self, h_a_alpha_out_dir, h_a_alpha_decomposition_T3_path, polsarpro_in_dir, is_trans_to_tif=True, is_read_to_dic=False, *args):
|
||||
|
|
|
|||
|
|
@ -50,7 +50,6 @@ else:
|
|||
DEBUG = False
|
||||
file =fileHandle(DEBUG)
|
||||
EXE_NAME = cf.get('exe_name')
|
||||
FILTER_SIZE = int(cf.get('filter_size'))
|
||||
soil_moisture_value_min = float(cf.get('product_value_min'))
|
||||
soil_moisture_value_max = float(cf.get('product_value_max'))
|
||||
pixelspace=float(cf.get('pixelspace'))
|
||||
|
|
|
|||
|
|
@ -376,6 +376,31 @@ class PreProcess:
|
|||
return {}
|
||||
return cutted_img_paths
|
||||
|
||||
def cut_imgs_VP(self, out_dir, para_names, processing_paras, shp_path, img_name):
|
||||
"""
|
||||
使用矢量数据裁剪影像
|
||||
:param para_names:需要检查的参数名称
|
||||
:param shp_path：裁剪的shp文件
|
||||
"""
|
||||
if len(para_names) == 0:
|
||||
return {}
|
||||
cutted_img_paths = {}
|
||||
try:
|
||||
for name in para_names:
|
||||
if name == 'Covering':
|
||||
img_name = img_name.split('_')[6] + '_'
|
||||
output_path = os.path.join(out_dir, img_name + name + '_cut.tif')
|
||||
else:
|
||||
output_path = os.path.join(out_dir, name + '_cut.tif')
|
||||
input_path = processing_paras[name]
|
||||
self.cut_img(output_path, input_path, shp_path)
|
||||
cutted_img_paths.update({name: output_path})
|
||||
logger.info('cut %s success!', name)
|
||||
except BaseException:
|
||||
logger.error('cut_img failed!')
|
||||
return {}
|
||||
return cutted_img_paths
|
||||
|
||||
@staticmethod
|
||||
def cut_img(output_path, input_path, shp_path):
|
||||
"""
|
||||
|
|
|
|||
|
|
@ -25,13 +25,14 @@ class BlockProcess:
|
|||
|
||||
@staticmethod
|
||||
def get_block_size(rows, cols):
|
||||
block_size = 512
|
||||
if (rows // 1024) <= 5 and (cols // 1024) <= 5:
|
||||
block_size = 512
|
||||
elif 5 < (rows // 1024) < 10 and 5 < (cols // 1024) < 10:
|
||||
block_size = 1024
|
||||
else:
|
||||
if rows <= 2048 or cols <= 2048:
|
||||
block_size = 512
|
||||
if (rows // 1024) <= 5 or (cols // 1024) <= 5:
|
||||
block_size = 1024
|
||||
elif 5 < (rows // 1024) and 5 < (cols // 1024):
|
||||
block_size = 2048
|
||||
|
||||
return block_size
|
||||
|
||||
# def get_block_size(rows, cols, block_size_config):
|
||||
|
|
|
|||
|
|
@ -70,7 +70,7 @@ class MachineLeaning:
|
|||
# 特征分块
|
||||
bp = BlockProcess()
|
||||
block_size = bp.get_block_size(rows, cols)
|
||||
|
||||
logger.info('block size is :%s', block_size)
|
||||
bp.cut(feature_tif_dir, workspace_block_tif_path, ['tif', 'tiff'], 'tif', block_size)
|
||||
img_dir, img_name = bp.get_file_names(workspace_block_tif_path, ['tif'])
|
||||
dir_dict_all = bp.get_same_img(img_dir, img_name)
|
||||
|
|
@ -209,7 +209,7 @@ class MachineLeaning:
|
|||
|
||||
band = ImageHandler.get_bands(path)
|
||||
if band == 1:
|
||||
features_array = np.zeros((1, 1024, 1024), dtype=float)
|
||||
features_array = np.zeros((1, block_size, block_size), dtype=float)
|
||||
feature_array = ImageHandler.get_data(path)
|
||||
features_array[0, :, :] = feature_array
|
||||
else:
|
||||
|
|
|
|||
|
|
@ -45,14 +45,14 @@ class DictXml:
|
|||
point_downright = [float(bottomRight.find("longitude").text), float(bottomRight.find("latitude").text)]
|
||||
scopes = [point_upleft, point_upright, point_downleft, point_downright]
|
||||
|
||||
point_upleft_buf = [float(topLeft.find("longitude").text) - 0.5, float(topLeft.find("latitude").text) + 0.5]
|
||||
point_upright_buf = [float(topRight.find("longitude").text) + 0.5, float(topRight.find("latitude").text) + 0.5]
|
||||
point_downleft_buf = [float(bottomLeft.find("longitude").text) - 0.5,
|
||||
float(bottomLeft.find("latitude").text) - 0.5]
|
||||
point_downright_buf = [float(bottomRight.find("longitude").text) + 0.5,
|
||||
float(bottomRight.find("latitude").text) - 0.5]
|
||||
scopes_buf = [point_upleft_buf, point_upright_buf, point_downleft_buf, point_downright_buf]
|
||||
return scopes
|
||||
point_upleft_buf = [float(topLeft.find("longitude").text) - 0.4, float(topLeft.find("latitude").text) + 0.4]
|
||||
point_upright_buf = [float(topRight.find("longitude").text) + 0.4, float(topRight.find("latitude").text) + 0.4]
|
||||
point_downleft_buf = [float(bottomLeft.find("longitude").text) - 0.4,
|
||||
float(bottomLeft.find("latitude").text) - 0.4]
|
||||
point_downright_buf = [float(bottomRight.find("longitude").text) + 0.4,
|
||||
float(bottomRight.find("latitude").text) - 0.4]
|
||||
scopes_buf = ([point_upleft_buf, point_upright_buf, point_downleft_buf, point_downright_buf], )
|
||||
return scopes, scopes_buf
|
||||
|
||||
|
||||
class xml_extend:
|
||||
|
|
|
|||
|
|
@ -37,7 +37,9 @@
|
|||
<ParaType>File</ParaType>
|
||||
<DataType>tar.gz</DataType>
|
||||
<ParaSource>Man</ParaSource>
|
||||
<ParaValue>F:\VegetationPhenology-likun\lijiang\GF3B_KSC_QPSI_007906_E100.2_N27.0_20230525_L1A_AHV_L10000190531-ortho.tar.gz</ParaValue>
|
||||
<ParaValue>
|
||||
F:\VegetationPhenology-likun\lijiang\GF3B_KSC_QPSI_007906_E100.2_N27.0_20230525_L1A_AHV_L10000190531-ortho.tar.gz
|
||||
</ParaValue>
|
||||
<MaxValue>DEFAULT</MaxValue>
|
||||
<MinValue>DEFAULT</MinValue>
|
||||
<OptionValue>DEFAULT</OptionValue>
|
||||
|
|
@ -107,13 +109,12 @@
|
|||
<ParaChsName>极化特征组合</ParaChsName>
|
||||
<Description>可选极化特征组合一、共14种特征（编号依次为0-13）
|
||||
Freeman：表面散射p_s(0)、偶次散射p_d(1)、体散射p_v(2);
|
||||
Touzi：散射角α_s(3)、散射相位ϕ_α(4)、目标散射对称度τ(5)、相对能量λ_i(6);
|
||||
Yamaguchi：表面散射f_s(7)、二次散射f_d(8)、体散射f_v(9)、螺旋体散射f_h(10);
|
||||
Cloude-Pottier：分解散射熵H(11)、反熵A(12)、平均散射角α(13)</Description>
|
||||
<ParaType>Value</ParaType>
|
||||
<DataType>string</DataType>
|
||||
<ParaSource>Man</ParaSource>
|
||||
<ParaValue>0,1,2,7,8,9,10</ParaValue>
|
||||
<ParaValue>0,1,2</ParaValue>
|
||||
<EnModification>True</EnModification>
|
||||
<EnMultipleChoice>True</EnMultipleChoice>
|
||||
<Control>UploadInput</Control>
|
||||
|
|
|
|||
|
|
@ -7,6 +7,7 @@
|
|||
@Date ：2021/9/6
|
||||
@Version ：1.0.0
|
||||
"""
|
||||
import csv
|
||||
import glob
|
||||
import logging
|
||||
import os
|
||||
|
|
@ -169,8 +170,8 @@ class PhenologyMain:
|
|||
# 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)
|
||||
cutted_img_paths = p.cut_imgs(self.__workspace_preprocessing_path, para_names_geo, self.__processing_paras,
|
||||
intersect_shp_path)
|
||||
cutted_img_paths = p.cut_imgs_VP(self.__workspace_preprocessing_path, para_names_geo, self.__processing_paras,
|
||||
intersect_shp_path, name)
|
||||
self.__preprocessed_paras.update({name + '_sim_ori': cutted_img_paths.get(name + '_sim_ori')})
|
||||
self.__preprocessed_paras.update({name + '_Covering': cutted_img_paths.get('Covering')})
|
||||
|
||||
|
|
@ -373,7 +374,9 @@ class PhenologyMain:
|
|||
"""
|
||||
# 生成每个时相的特征， 并提取训练集和测试集
|
||||
# 每个时相的影像生成特征图
|
||||
X_train, Y_train = None, None
|
||||
|
||||
X_train = []
|
||||
Y_train = []
|
||||
flag = True
|
||||
total_name_list = []
|
||||
X_test_dic = {}
|
||||
|
|
@ -381,35 +384,45 @@ class PhenologyMain:
|
|||
|
||||
feature_dir, train_data_dic = self.create_feature_single_tar(name)
|
||||
# 生成训练集
|
||||
X_train_part, Y_train_part, optimal_feature = ml.gene_optimal_train_set(train_data_dic, feature_dir, 0.08, 0.85)
|
||||
# X_train_part, Y_train_part, optimal_feature = ml.gene_optimal_train_set(train_data_dic, feature_dir, 0.08, 0.7)
|
||||
X_train_part, Y_train_part = ml.gene_train_set(train_data_dic, feature_dir)
|
||||
name_list = ml.get_name_list(feature_dir)
|
||||
if len(optimal_feature) <= 0:
|
||||
logger.error('特征筛选结果为空，无可用特征作为训练集')
|
||||
continue
|
||||
|
||||
# 生成测试集合
|
||||
X_test_path_list = ml.gene_test_set(feature_dir, optimal_feature)
|
||||
# X_test_path_list = ml.gene_test_set(feature_dir, optimal_feature)
|
||||
|
||||
X_test_dic.update({name: X_test_path_list})
|
||||
# X_test_dic.update({name: X_test_path_list})
|
||||
X_test_dic.update({name + '_features': feature_dir})
|
||||
|
||||
if X_train == []:
|
||||
X_train = X_train_part
|
||||
Y_train = Y_train_part
|
||||
total_name_list = name_list
|
||||
else:
|
||||
X_train = np.vstack((X_train, X_train_part))
|
||||
Y_train = np.hstack((Y_train, Y_train_part))
|
||||
total_name_list = total_name_list + name_list
|
||||
|
||||
logger.info("generate train and test set success!")
|
||||
logger.info('progress bar: 30%')
|
||||
|
||||
optimal_X_train, optimal_Y_train, optimal_feature = ml.sel_optimal_feature(X_train, Y_train, total_name_list, correlation_threshold=0.7)
|
||||
|
||||
# RF
|
||||
clf = ml.trainRF(X_train, Y_train)
|
||||
clf = ml.trainRF(optimal_X_train, optimal_Y_train)
|
||||
logger.info('RF train success!')
|
||||
logger.info('progress bar: 80%')
|
||||
|
||||
# 测试数据
|
||||
logger.info('mode testing')
|
||||
|
||||
for name in self.__processing_paras['name_list']:
|
||||
feature_dir = X_test_dic.get(name + '_features')
|
||||
X_test_path_list = ml.gene_test_set(feature_dir, optimal_feature)
|
||||
in_tif_paths = list(glob.glob(os.path.join(feature_dir, '*.tif')))
|
||||
rows = ImageHandler.get_img_height(in_tif_paths[0])
|
||||
cols = ImageHandler.get_img_width(in_tif_paths[0])
|
||||
proj_geo, geo_geo, cover_data_geo = self.imageHandler.read_img(in_tif_paths[0])
|
||||
product_path = ml.predict_VP(clf, X_test_path_list, name, self.__workspace_processing_path, rows, cols)
|
||||
# product_path = ml.predict_VP(clf, X_test_dic.get(name), name, self.__workspace_processing_path, rows, cols)
|
||||
|
||||
proj, geo, cover_data = self.imageHandler.read_img(product_path)
|
||||
# 形态学（闭运算）去roi区域噪点
|
||||
|
|
@ -417,7 +430,8 @@ class PhenologyMain:
|
|||
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']):
|
||||
train_dic = self.get_train_dic(self.__processing_paras['MeasuredData'])
|
||||
for id, class_id in zip(train_dic['ids'], train_dic['class_ids']):
|
||||
cover_data[np.where(cover_data == id)] = class_id
|
||||
roi_img = self.imageHandler.get_band_array(self.create_roi(in_tif_paths[0]))
|
||||
# 获取影像roi区域
|
||||
|
|
@ -443,64 +457,11 @@ class PhenologyMain:
|
|||
# CreateProductXml(para_dict, model_path, out_xml).create_standard_xml()
|
||||
shutil.copy(meta_xml_path, out_xml)
|
||||
logger.info('mode test success!')
|
||||
logger.info('progress bar: 80%')
|
||||
# 文件夹打包
|
||||
file.make_targz(self.__out_para, self.__product_dic)
|
||||
logger.info('progress bar: 100%')
|
||||
|
||||
# """
|
||||
# 算法主处理函数
|
||||
# :return: True or False
|
||||
# """
|
||||
# # 生成每个时相的特征， 并提取训练集和测试集
|
||||
# # 每个时相的影像生成特征图
|
||||
# X_train, Y_train = None, None
|
||||
# flag = True
|
||||
# total_name_list = []
|
||||
# test_data = []
|
||||
# X_test_dic = {}
|
||||
# for name in self.__processing_paras['name_list']:
|
||||
# feature_dir, train_data_dic, test_data_part, type_map = self.create_feature_single_tar(name)
|
||||
# #生成训练集
|
||||
# X_train_part, Y_train_part = ml.gene_train_set(train_data_dic, feature_dir)
|
||||
# name_list = ml.get_name_list(feature_dir)
|
||||
# # 生成测试集合
|
||||
# rows, cols = self.get_name_rows_cols(name)
|
||||
# name_featuresPath_dic_part = ml.vegetationPhenology_combine_feature(feature_dir, self.__workspace_processing_path, name, rows, cols, DEBUG)
|
||||
# X_test_dic_part = self.gene_test_set(test_data_part, name_featuresPath_dic_part, name)
|
||||
# X_test_dic.update(X_test_dic_part)
|
||||
# if flag:
|
||||
# X_train = X_train_part
|
||||
# Y_train = Y_train_part
|
||||
# total_name_list = name_list
|
||||
# flag = False
|
||||
# test_data = test_data_part
|
||||
# else:
|
||||
# X_train = np.vstack((X_train, X_train_part))
|
||||
# Y_train = np.hstack((Y_train, Y_train_part))
|
||||
# total_name_list = total_name_list + name_list
|
||||
# test_data = test_data + test_data_part
|
||||
#
|
||||
# logger.info("create_features success!")
|
||||
# logger.info('progress bar: 20%')
|
||||
#
|
||||
# optimal_X_train, optimal_Y_train, optimal_feature = ml.sel_optimal_feature(X_train, Y_train, total_name_list, correlation_threshold=0.7)
|
||||
#
|
||||
# logger.info("generate train and test set success!")
|
||||
# logger.info('progress bar: 30%')
|
||||
#
|
||||
# #RF
|
||||
# clf = ml.trainRF(optimal_X_train, optimal_Y_train)
|
||||
# logger.info('svm train success!')
|
||||
# logger.info('progress bar: 80%')
|
||||
#
|
||||
# # 测试数据
|
||||
# logger.info('mode testing')
|
||||
# product_path = self.predict(clf, X_test_dic, optimal_feature, type_map, start)
|
||||
# logger.info('mode test success!')
|
||||
# self.create_meta_file(product_path)
|
||||
# # 文件夹打包
|
||||
# file.make_targz(self.__out_para, self.__product_dic)
|
||||
# logger.info('progress bar: 100%')
|
||||
|
||||
def predict(self, mode, X_test_dic, validity_list, type_map, start):
|
||||
# 测试数据
|
||||
|
|
@ -558,6 +519,23 @@ class PhenologyMain:
|
|||
|
||||
return product_geo_path
|
||||
|
||||
def get_train_dic(self, csv_path):
|
||||
reader = csv.reader(open(csv_path, newline=''))
|
||||
ids = []
|
||||
class_ids = []
|
||||
csv_list = []
|
||||
for line_data in reader:
|
||||
csv_list.append(line_data)
|
||||
for data in csv_list[1:]:
|
||||
ids.append(data[0])
|
||||
class_ids.append(data[1])
|
||||
|
||||
train_data_dic = {}
|
||||
train_data_dic.update({"ids": ids})
|
||||
train_data_dic.update({"class_ids": class_ids})
|
||||
|
||||
return train_data_dic
|
||||
|
||||
def resampleImgs(self, name, refer_img_path):
|
||||
|
||||
cover_rampling_path = os.path.join(self.__workspace_processing_path, name + "_cover.tif")
|
||||
|
|
|
|||
Loading…
Reference in New Issue