import pyrat
import numpy as np
class SVA(pyrat.FilterWorker):
"""
Spatial Variant Apodisation (SVA) filter: Removes SINC function sidelobes
Author: J. Fischer, DLR/HR
Date: 24/10/2014
Docu: Fischer,Pupeza,Scheiber: Sidelobe Suppression Using the SVA Method for SAR Images and Sounding Radars, EUSAR Proceedings, 2006
"""
# Comment: Needs reworking before getting accepted for the GUI
# gui = {'menu': 'SAR|Sidelobe control', 'entry': 'Spatial Variant Apodization (SVA)'}
para = [
{'var': 'ov', 'value': [2, 2], 'type': 'int', 'range': [1, 10], 'text': 'Desired Oversampling', 'subtext': ['azimuth', 'range']}
]
def __init__(self, *args, **kwargs):
super(SVA, self).__init__(*args, **kwargs)
self.name = "SVA FILTER"
if 'ov' not in self.__dict__:
self.ov = [2, 2]
elif isinstance(self.ov, int):
self.ov = [self.ov] * 2
self.blockprocess = True
self.blockoverlap = self.ov[0] # = 1 pixel margin x azimuth oversampling factor
def pre(self, *args, **kwargs):
#pyrat.filter.unweight(ov=self.ov)
pass
def post(self, *args, **kwargs):
pass
def filter(self, array, *args, **kwargs):
# 1-D Real Arrays
if array.ndim == 1 and np.isrealobj(array):
return self.svafilter(array, self.ov)
# 1-D Complex Arrays
if array.ndim == 1 and np.iscomplexobj(array):
return self.sva1D(array, self.ov)
# 2-D Complex Arrays
if array.ndim == 2 and np.iscomplexobj(array):
return self.sva2D(array, self.ov)
# 3-D Complex Arrays
if array.ndim == 3 and np.iscomplexobj(array):
p = array.shape
for k in range(0,p[0]):
array[k,:,:] = self.sva2D(array[k,:,:], self.ov)
return array
else:
print(" ERROR: Bad input.")
return None
def sva2D(self, image0, ov):
# SVA for complex valued 2D arrays
print(" Applying SVA with ov=["+str(ov[0])+","+str(ov[1])+"] ... ")
s = image0.shape
image1 = np.empty((s[0],s[1]), dtype='complex64')
image2 = np.empty((s[0],s[1]), dtype='complex64')
for k in range(0,s[0]):
image1[k,:] = self.sva1D( image0[k,:], ov[1] ) # SVA in range
for k in range(0,s[1]):
image2[:,k] = self.sva1D( image1[:,k], ov[0] ) # SVA in azimuth
print(" Applying SVA with ov=["+str(ov[0])+","+str(ov[1])+"] done. ")
return image2
def sva1D(self, image, ov):
# SVA for complex valued 1D arrays
#k,d = get_kd(image) # Note that n = k*d
k = image.size / ov
d = ov
# n = total signal length
# k = signal bandwidth in [pixel], k = supp(spectrum), k is the 'actual signal length'
# d = integer oversampling factor = number of pixels between sinc zero crossings, d is the redundancy factor
return self.svafilter(image.real,k,d) + 1j * self.svafilter(image.imag,k,d)
def svafilter(self, image,k,d):
# SVA for real valued 1D arrays
image_sva = np.empty(image.size,float)
ind = np.arange(0,k) * d
for off in range(0,d): # run through all "offsets" (between zero crossings)
image_sva[ind+off] = self.svafilterkernel( image[ind+off], k)
return image_sva
def svafilterkernel(self, sig,k):
# SVA filter kernel for real valued 1D arrays
np.seterr(divide='ignore',invalid='ignore')
wu = np.zeros(k)
for m in range(1,k-1):
wu[m]=-sig[m]/(sig[m-1]+sig[m+1])
np.seterr(divide='warn',invalid='warn')
a = np.zeros(k)
for m in range(1,k-1):
if wu[m] <= 0:
a[m] = sig[m] # mainlobe (preserved)
elif (wu[m] >= 0.5): #sidelobe (0 < wu[m] < 0.5) set to zero
a[m]=sig[m]+0.5*(sig[m-1]+sig[m+1]) # mixed signal (nothing done)
return a
@pyrat.docstringfrom(SVA)
def sva(*args, **kwargs):
return SVA(*args, **kwargs).run(**kwargs)
