Files
BRDF/Flexbrdf/hytools/io/envi.py
2026-04-10 16:46:45 +08:00

698 lines
22 KiB
Python

# -*- coding: utf-8 -*-
"""
HyTools: Hyperspectral image processing library
Copyright (C) 2021 University of Wisconsin
Authors: Adam Chlus, Zhiwei Ye, Philip Townsend.
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, version 3 of the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <https://www.gnu.org/licenses/>.
Functions for reading and writing ENVI formatted binary files
Todo:
* Implement opening of ENVI files with different byte order
"""
import os
import sys
from collections import Counter
import numpy as np
# ENVI datatype conversion dictionary
dtype_dict = {1:np.uint8,
2:np.int16,
3:np.int32,
4:np.float32,
5:np.float64,
12:np.uint16,
13:np.uint32,
14:np.int64,
15:np.uint64}
# Dictionary of all ENVI header fields
field_dict = {"acquisition time": "str",
"band names":"list_str",
"bands": "int",
"bbl": "list_float",
"byte order": "int",
"class lookup": "str",
"class names": "str",
"classes": "int",
"cloud cover": "float",
"complex function": "str",
"coordinate system string": "str",
"correction factors": "list_float",
"data gain values": "list_float",
"data ignore value": "float",
"data offset values": "list_float",
"data reflectance gain values": "list_float",
"data reflectance offset values": "list_float",
"data type": "int",
"default bands": "list_float",
"default stretch": "str",
"dem band": "int",
"dem file": "str",
"description": "str",
"envi description":"str",
"file type": "str",
"fwhm": "list_float",
"geo points": "list_float",
"header offset": "int",
"interleave": "str",
"lines": "int",
"map info": "list_str",
"pixel size": "list_str",
"projection info": "str",
"read procedures": "str",
"reflectance scale factor": "float",
"rpc info": "str",
"samples":"int",
"security tag": "str",
"sensor type": "str",
"smoothing factors": "list_float",
"solar irradiance": "float",
"spectra names": "list_str",
"sun azimuth": "float",
"sun elevation": "float",
"wavelength": "list_float",
"wavelength units": "str",
"x start": "float",
"y start": "float",
"z plot average": "str",
"z plot range": "str",
"z plot titles": "str"}
def open_envi(hy_obj,anc_path = {}, ext = False, glt_path = None):
"""Open ENVI formatted image file and populate Hytools object.
Args:
src_file (str): Pathname of input ENVI image file, header assumed to be located in
same directory.
anc_path (dict): Dictionary with pathnames and band numbers of ancillary datasets.
ext: (bool) Input ENVI file has a file extension
Returns:
HyTools file object: Populated HyTools file object.
"""
header_file = os.path.splitext(hy_obj.file_name)[0] + ".hdr"
if not os.path.isfile(header_file):
print("ERROR: Header file not found.")
return None
header_dict = parse_envi_header(header_file)
hy_obj.lines = header_dict["lines"]
hy_obj.columns = header_dict["samples"]
hy_obj.bands = header_dict["bands"]
hy_obj.bad_bands = np.array([False for band in range(hy_obj.bands)])
hy_obj.interleave = header_dict["interleave"]
hy_obj.fwhm = header_dict["fwhm"]
hy_obj.wavelengths = header_dict["wavelength"]
hy_obj.wavelength_units = header_dict["wavelength units"]
hy_obj.dtype = dtype_dict[header_dict["data type"]]
hy_obj.no_data = header_dict['data ignore value']
hy_obj.map_info = header_dict['map info']
hy_obj.byte_order = header_dict['byte order']
hy_obj.anc_path = anc_path
hy_obj.header_file = header_file
hy_obj.transform = calc_geotransform(header_dict['map info'])
if bool(header_dict['coordinate system string']):
hy_obj.projection = header_dict['coordinate system string']
else:
hy_obj.projection = ''
if hy_obj.byte_order == 1:
hy_obj.endianness = 'big'
else:
hy_obj.endianness = 'little'
if isinstance(header_dict['bbl'],np.ndarray):
hy_obj.bad_bands = np.array([x==1 for x in header_dict['bbl']])
if header_dict["interleave"] == 'bip':
hy_obj.shape = (hy_obj.lines, hy_obj.columns, hy_obj.bands)
elif header_dict["interleave"] == 'bil':
hy_obj.shape = (hy_obj.lines, hy_obj.bands, hy_obj.columns)
elif header_dict["interleave"] == 'bsq':
hy_obj.shape = (hy_obj.bands, hy_obj.lines, hy_obj.columns)
else:
print("ERROR: Unrecognized interleave type.")
hy_obj = None
# If no_data value is not specified guess using image corners.
if hy_obj.no_data is None:
hy_obj.load_data()
band_ind = 5 if hy_obj.bands > 10 else 0
if header_dict["interleave"] == 'bip':
up_l = hy_obj.data[0,0,band_ind]
up_r = hy_obj.data[0,-1,band_ind]
low_l = hy_obj.data[-1,0,band_ind]
low_r = hy_obj.data[-1,-1,band_ind]
elif header_dict["interleave"] == 'bil':
up_l = hy_obj.data[0,band_ind,0]
up_r = hy_obj.data[0,band_ind,-1]
low_l = hy_obj.data[-1,band_ind,0]
low_r = hy_obj.data[-1,band_ind,-1]
elif header_dict["interleave"] == 'bsq':
up_l = hy_obj.data[band_ind,0,0]
up_r = hy_obj.data[band_ind,0,-1]
low_l = hy_obj.data[band_ind,-1,0]
low_r = hy_obj.data[band_ind,-1,-1]
if hy_obj.endianness != sys.byteorder:
up_l = up_l.byteswap()
up_r = up_r.byteswap()
low_l = low_l.byteswap()
low_r = low_r.byteswap()
counts = {v: k for k, v in Counter([up_l,up_r,low_l,low_r]).items()}
hy_obj.no_data = counts[max(counts.keys())]
hy_obj.close_data()
if bool(glt_path):
glt_meta_dict = parse_glt_envi(glt_path)
hy_obj.glt_path = glt_meta_dict["glt_path"]
hy_obj.glt_map_info = glt_meta_dict["map_info"]
hy_obj.lines_glt = glt_meta_dict["lines_glt"]
hy_obj.columns_glt = glt_meta_dict["columns_glt"]
hy_obj.glt_transform = glt_meta_dict["transform"]
hy_obj.glt_projection = glt_meta_dict["projection"]
del glt_meta_dict
del header_dict
return hy_obj
class WriteENVI:
"""Iterator class for writing to an ENVI data file.
"""
def __init__(self,output_name,header_dict):
"""
Args:
output_name (str): Pathname of output ENVI data file.
header_dict (dict): Dictionary containing ENVI header information.
Returns:
None.
"""
self.interleave = header_dict['interleave']
self.header_dict = header_dict
self.output_name =output_name
dtype = dtype_dict[header_dict["data type"]]
lines = header_dict['lines']
columns = header_dict['samples']
bands = header_dict['bands']
if self.interleave == "bip":
self.data = np.memmap(output_name,dtype = dtype,
mode='w+', shape = (lines,columns,bands))
elif self.interleave == "bil":
self.data = np.memmap(output_name,dtype = dtype,
mode='w+', shape =(lines,bands,columns))
elif self.interleave == "bsq":
self.data = np.memmap(output_name,dtype = dtype,
mode='w+',shape =(bands,lines,columns))
write_envi_header(self.output_name,self.header_dict)
def write_line(self,line,index):
"""
Args:
line (numpy.ndarray): Line array (columns,bands).
index (int): Zero-based line index.
Returns:
None.
"""
if self.interleave == "bip":
self.data[index,:,:] = line
elif self.interleave == "bil":
self.data[index,:,:] = np.moveaxis(line,0,1)
elif self.interleave == "bsq":
self.data[:,index,:] = np.moveaxis(line,0,1)
def write_line_glt(self,arr,glt_indices_y,glt_indices_x):
"""
Args:
line (numpy.ndarray): Line array (columns,bands).
index (int): Zero-based line index.
Returns:
None.
"""
if self.interleave == "bip":
self.data[glt_indices_y,glt_indices_x,:] = arr
elif self.interleave == "bil":
self.data[glt_indices_y,:,glt_indices_x] = arr #np.moveaxis(line,0,1)
elif self.interleave == "bsq":
self.data[:,glt_indices_y,glt_indices_x] = np.moveaxis(arr,0,1)
def write_column(self,column,index):
"""
Args:
column (numpy.ndarray): Column array (lines,bands).
index (int): Zero-based column index.
Returns:
None.
"""
if self.interleave == "bip":
self.data[:,index,:] = column
elif self.interleave == "bil":
self.data[:,:,index] = column
elif self.interleave == "bsq":
self.data[:,:,index] = np.moveaxis(column,0,1)
def write_band(self,band,index):
"""
Args:
band (numpy.ndarray): Band array (lines,columns).
index (int): Zero-based band index.
Returns:
None.
"""
if self.interleave == "bip":
self.data[:,:,index] = band
elif self.interleave == "bil":
self.data[:,index,:] = band
elif self.interleave == "bsq":
self.data[index,:,:]= band
def write_band_glt(self,band,index,glt_indices,fill_mask):
"""
Args:
band (numpy.ndarray): Band array (lines,columns).
index (int): Zero-based band index.
glt_indices (numpy.ndarray,numpy.ndarray): Zero-based tuple indices.
Returns:
None.
"""
if self.interleave == "bip":
self.data[:,:,index][fill_mask] = band[glt_indices]
self.data[:,:,index][~fill_mask] = self.header_dict['data ignore value']
elif self.interleave == "bil":
self.data[:,index,:][fill_mask] = band[glt_indices]
self.data[:,index,:][~fill_mask] = self.header_dict['data ignore value']
elif self.interleave == "bsq":
self.data[index,:,:][fill_mask] = band[glt_indices]
self.data[index,:,:][~fill_mask] = self.header_dict['data ignore value']
def write_chunk(self,chunk,line_index,column_index):
"""
Args:
chunk (TYPE): Chunks array (chunk lines,chunk columns,bands).
line_index (int): Zero-based upper line index.
column_index (int): Zero-based left column index.
Returns:
None.
"""
x_start = column_index
x_end = column_index + chunk.shape[1]
y_start = line_index
y_end = line_index + chunk.shape[0]
if self.interleave == "bip":
self.data[y_start:y_end,x_start:x_end,:] = chunk
elif self.interleave == "bil":
self.data[y_start:y_end,:,x_start:x_end] = np.moveaxis(chunk,-1,-2)
elif self.interleave == "bsq":
self.data[:,y_start:y_end,x_start:x_end] = np.moveaxis(chunk,-1,0)
def write_pixel(self,pixel,line_index,column_index):
"""
Args:
pixel (TYPE): pixel array (bands).
line_index (int): Zero-based upper line index.
column_index (int): Zero-based left column index.
Returns:
None.
"""
if self.interleave == "bip":
self.data[line_index,column_index,:] = pixel
elif self.interleave == "bil":
self.data[line_index,:,column_index] = pixel
elif self.interleave == "bsq":
self.data[:,line_index,column_index] = pixel
def close(self):
"""Delete numpy memmap.
"""
del self.data
def envi_header_from_neon(hy_obj, interleave = 'bsq'):
"""Create an ENVI header dictionary from NEON metadata
Args:
hy_obj (Hytools object): Populated HyTools file object.
interleave (str, optional): Date interleave type. Defaults to 'bil'.
Returns:
dict: Populated ENVI header dictionary.
"""
header_dict = {}
header_dict["ENVI description"] = "{}"
header_dict["samples"] = hy_obj.columns
header_dict["lines"] = hy_obj.lines
header_dict["bands"] = hy_obj.bands
header_dict["header offset"] = 0
header_dict["file type"] = "ENVI Standard"
header_dict["data type"] = 2
header_dict["interleave"] = interleave
header_dict["sensor type"] = ""
header_dict["byte order"] = 0
header_dict["map info"] = hy_obj.map_info
header_dict["coordinate system string"] = hy_obj.projection
header_dict["wavelength units"] = hy_obj.wavelength_units
header_dict["data ignore value"] =hy_obj.no_data
header_dict["wavelength"] =hy_obj.wavelengths
return header_dict
def envi_header_from_nc(hy_obj, interleave = 'bsq', warp_glt = False):
"""Create an ENVI header dictionary from NetCDF metadata
Args:
hy_obj (Hytools object): Populated HyTools file object.
interleave (str, optional): Date interleave type. Defaults to 'bil'.
Returns:
dict: Populated ENVI header dictionary.
"""
header_dict = {}
header_dict["ENVI description"] = "{}"
if warp_glt == False:
header_dict["samples"] = hy_obj.columns
header_dict["lines"] = hy_obj.lines
header_dict["map info"] = hy_obj.map_info
header_dict["coordinate system string"] = "{%s}" % hy_obj.projection if hy_obj.projection else "{}"
header_dict["projection"] = hy_obj.projection
header_dict["transform"] = hy_obj.transform
else:
header_dict["samples"] = hy_obj.columns_glt
header_dict["lines"] = hy_obj.lines_glt
header_dict["map info"] = hy_obj.glt_map_info
header_dict["coordinate system string"] = "{%s}" % hy_obj.glt_projection if hy_obj.glt_projection else "{}"
header_dict["projection"] = hy_obj.glt_projection
header_dict["transform"] = hy_obj.glt_transform
header_dict["bands"] = 2 #hy_obj.bands
header_dict["header offset"] = 0
header_dict["file type"] = "ENVI Standard"
header_dict["data type"] = 4
header_dict["interleave"] = interleave
header_dict["sensor type"] = ""
header_dict["byte order"] = 0
header_dict["wavelength units"] = hy_obj.wavelength_units
header_dict["data ignore value"] = hy_obj.no_data
header_dict["wavelength"] = hy_obj.wavelengths
return header_dict
def write_envi_header(output_name,header_dict,mode = 'w'):
"""Write ENVI header file to disk.
Args:
output_name (str): Header file pathname.
header_dict (dict): Populated ENVI header dictionary.
mode (str): File open mode. default: w
Returns:
None.
"""
base_name = os.path.splitext(output_name)[0]
header_file = open(base_name + ".hdr",mode)
header_file.write("ENVI\n")
for key in header_dict.keys():
value = header_dict[key]
# Convert list to comma separated strings
if isinstance(value,(list,np.ndarray)):
value = "{%s}" % ",".join(map(str, value))
elif key == "coordinate system string" and value and isinstance(value, str):
# 对 coordinate system string 字段确保有花括号包围
if not value.startswith("{"):
value = "{%s}" % value
else:
value = str(value)
# Skip entires with nan as value
if value != 'None':
header_file.write("%s = %s\n" % (key,value))
header_file.close()
def envi_header_dict():
"""
Returns:
dict: Empty ENVI header dictionary.
"""
return {key:None for (key,value) in field_dict.items()}
def envi_read_line(data,index,interleave):
"""
Args:
data (numpy.memmap): Numpy memory-map.
index (int): Zero-based line index.
interleave (str): Data interleave type.
Returns:
numpy.ndarray: Line array (columns, bands).
"""
if interleave == "bip":
line = data[index,:,:]
elif interleave == "bil":
line = np.moveaxis(data[index,:,:],0,1)
elif interleave == "bsq":
line = np.moveaxis(data[:,index,:],0,1)
return line
def envi_read_column(data,index,interleave):
"""
Args:
data (numpy.memmap): Numpy memory-map.
index (int): Zero-based column index.
interleave (str): Data interleave type.
Returns:
numpy.ndarray: Column array (lines,bands).
"""
if interleave == "bip":
column = data[:,index,:]
elif interleave == "bil":
column = data[:,:,index]
elif interleave == "bsq":
column = np.moveaxis(data[:,:,index],0,1)
return column
def envi_read_band(data,index,interleave):
"""
Args:
data (numpy.memmap): Numpy memory-map.
index (int): Zero-based line index.
interleave (str): Data interleave type.
Returns:
numpy.ndarray: Band array (lines,columns).
"""
if interleave == "bip":
band = data[:,:,index]
elif interleave == "bil":
band = data[:,index,:]
elif interleave == "bsq":
band = data[index,:,:]
return band
def envi_read_pixels(data,lines,columns,interleave):
"""
Args:
data (numpy.memmap): Numpy memory-map.
lines (list): List of zero-indexed line indices.
columns (list): List of zero-indexed column indices.
interleave (str): Data interleave type.
Returns:
numpy.ndarray: Pixel array (pixels,bands).
"""
if interleave == "bip":
pixels = data[lines,columns,:]
elif interleave == "bil":
pixels = data[lines,:,columns]
elif interleave == "bsq":
pixels = data[:,lines,columns]
return pixels
def envi_read_chunk(data,col_start,col_end,line_start,line_end,interleave):
"""
Args:
data (numpy.memmap): Numpy memory-map.
col_start (int): Zero-based left column index.
col_end (int): Non-inclusive zero-based right column index.
line_start (int): Zero-based top line index.
line_end (int): Non-inclusive zero-based bottom line index.
interleave (str): Data interleave type.
Returns:
numpy.ndarray: Chunk array (line_end-line_start,col_end-col_start,bands).
"""
if interleave == "bip":
chunk = data[line_start:line_end,col_start:col_end,:]
elif interleave == "bil":
chunk = np.moveaxis(data[line_start:line_end,:,col_start:col_end],-1,-2)
elif interleave == "bsq":
chunk = np.moveaxis(data[:,line_start:line_end,col_start:col_end],0,-1)
return chunk
def calc_geotransform(mapinfo):
if mapinfo[-1].startswith('rotation'):
rot_ang_rad = np.radians(float(mapinfo[-1].split('=')[1]))
pixel_size = float(mapinfo[5])
new_rot_mat = pixel_size * np.array([[np.cos(rot_ang_rad),-np.sin(rot_ang_rad)],[np.sin(rot_ang_rad),np.cos(rot_ang_rad)]])@np.array([[1,0],[0,-1]])
geotransform = (float(mapinfo[3]),new_rot_mat[0,0],new_rot_mat[0,1],
float(mapinfo[4]),new_rot_mat[1,0],new_rot_mat[1,1])
else:
# same as 0 rotation
geotransform = (float(mapinfo[3]),float(mapinfo[5]),0,
float(mapinfo[4]),0,-float(mapinfo[6]))
return geotransform
def parse_glt_envi(glt_path):
glt_meta_dict = {}
glt_meta_dict["glt_path"] = glt_path
glt_header_file = os.path.splitext(glt_path[list(glt_path.keys())[0]][0])[0] + ".hdr"
glt_header=parse_envi_header(glt_header_file)
glt_meta_dict["map_info"] = glt_header["map info"]
glt_meta_dict["lines_glt"] = glt_header["lines"]
glt_meta_dict["columns_glt"] = glt_header["samples"]
glt_meta_dict["transform"] = calc_geotransform(glt_header["map info"])
if "coordinate system string" in glt_header:
glt_meta_dict["projection"] = glt_header["coordinate system string"]
else:
glt_meta_dict["projection"] = ''
return glt_meta_dict
def parse_envi_header(header_file):
"""
Args:
header_file (str): Header file pathname.
Returns:
dict: Populated header dictionary.
"""
header_dict = envi_header_dict()
header_file = open(header_file,'r')
line = header_file.readline()
while line :
if "=" in line:
key,value = line.rstrip().split("=",1)
# Add fields not in ENVI default list
if key.strip() not in field_dict.keys():
field_dict[key.strip()] = "str"
val_type = field_dict[key.strip()]
if "{" in value and not "}" in value:
while "}" not in line:
line = header_file.readline()
value+=line
if '{}' in value:
value = None
elif val_type == "list_float":
value= np.array([float(x) for x in value.translate(str.maketrans("\n{}"," ")).split(",")])
elif val_type == "list_int":
value= np.array([int(x) for x in value.translate(str.maketrans("\n{}"," ")).split(",")])
elif val_type == "list_str":
value= [x.strip() for x in value.translate(str.maketrans("\n{}"," ")).split(",")]
elif val_type == "int":
value = int(value.translate(str.maketrans("\n{}"," ")))
elif val_type == "float":
value = float(value.translate(str.maketrans("\n{}"," ")))
elif val_type == "str":
value = value.translate(str.maketrans("\n{}"," ")).strip().lower()
header_dict[key.strip()] = value
line = header_file.readline()
# Fill unused fields with None
for key in field_dict:
if key not in header_dict.keys():
header_dict[key] = None
header_file.close()
return header_dict