Script for masking known sources using the Gaia database.

tools/masking.py

@@ -0,0 +1,189 @@
+import os
+import astropy.units as u
+import numpy as np
+import numpy.ma as ma
+import math
+import sys
+import matplotlib.pyplot as plt
+from photutils import centroid_sources
+from astropy.wcs import WCS
+from astropy.io import fits
+from astropy.coordinates import SkyCoord
+from regions import CirclePixelRegion, PixCoord
+from astropy.wcs.utils import skycoord_to_pixel, pixel_to_skycoord
+from astropy.coordinates import ICRS,  FK5, SkyCoord
+from astroquery.gaia import Gaia
+from matplotlib.colors import LogNorm
+CRVAL1 =[]
+CRVAL2 = []
+CD1_1 =[]
+CD1_2 =[]
+CD2_1 =[]
+CD2_2 =[]
+VALS = []
+Big_Fudge = 1.5 #Arbitrary constant to scale large, bright sources. 
+small_fudge = .2 #Arbitrary constant to sclae smaller, dim sources.
+
+def get_wcs(images): #extracts necessary WCS info from science images.
+    L = int(len(images))
+    weight = np.ones(L)
+    weight[0] = 10
+    for W in images:
+        CRVAL1.append(W.header['CRVAL1'])
+        CRVAL2.append(W.header['CRVAL2'])
+        CD1_1.append(W.header['CD1_1'])
+        CD1_2.append(W.header['CD1_2'])
+        CD2_1.append(W.header['CD2_1'])
+        CD2_2.append(W.header['CD2_2'])
+    NAXIS = images[0].header['NAXIS']
+    NAXIS1 = images[0].header['NAXIS1']
+    NAXIS2 = images[0].header['NAXIS2']
+    CRPIX1 = images[0].header['CRPIX1']
+    CRPIX2 = images[0].header['CRPIX2']
+    CTYPE1 = images[0].header['CTYPE1']
+    CTYPE2 = images[0].header['CTYPE2']
+    CUNIT1 = images[0].header['CUNIT1']
+    CUNIT2 = images[0].header['CUNIT2'] 
+    VAL1 = np.average(CRVAL1, weights=weight) #need to add weighting back into this but letting it go until can get things working again.
+    VAL2 = np.average(CRVAL2, weights=weight)
+    VALS.append(VAL1)
+    VALS.append(VAL2)
+    CD11 = np.average(CD1_1, weights=weight)
+    CD12 = np.average(CD1_2, weights=weight)
+    CD21 = np.average(CD2_1, weights=weight)
+    CD22 = np.average(CD2_2, weights=weight)
+    total_wcs = WCS(naxis = int(NAXIS))
+    total_wcs.wcs.crpix = [CRPIX1, CRPIX2]
+    total_wcs.wcs.crval = [VAL1, VAL2]
+    total_wcs.wcs.cunit = [CUNIT1, CUNIT2]
+    total_wcs.wcs.ctype = [CTYPE1, CTYPE2]
+    total_wcs.wcs.cd = [(CD11, CD12), (CD21, CD22)]
+    total_wcs.array_shape = [NAXIS2, NAXIS1]
+    return total_wcs
+
+
+
+def reference(im): #runs GAIA cone search
+   ra = VALS[0]
+   dec = VALS[1]
+   coord = SkyCoord(ra, dec, unit = (u.deg, u.deg), frame = 'icrs')
+   radius = u.Quantity(.3, u.deg) #.3 deg is good enough for normal sized images can be adjusted.
+   Gaia.ROW_LIMIT = -1 # -1 returns all objects in search.
+   j = Gaia.cone_search_async(coord, radius)
+   r = j.get_results()
+   print('done searching')
+   return r
+
+
+
+def mask(image, scis): # WILL REQUIRE TWO ARGS: image to be masked (i.e. template image) and list of science images passed from main.
+    data = image
+    w = get_wcs(scis)
+    empty_mask = ma.zeros([data.shape[0]+100, data.shape[1]+100]) #adds padding to the image so that all sources in the frame can be masked
+    #print(w)
+    f = reference(image)
+    for source in f:
+        if  source['ra_error'] > .3 and  source['dec_error'] > .3: #filters out ra  &  dec error greater than .3 degrees
+            continue
+        print('STEP1')
+        if source['phot_g_mean_mag'] >18 and source['phot_bp_mean_mag']  > 18 and source['phot_rp_mean_mag'] > 18:
+            continue #filters  out sources with  RGB mag  greater than 18 -> sources dimmer than mag 18 (should be or?)
+        print('STEP2')
+        point = SkyCoord(ra = source['ra'],  dec= source['dec'], unit = (u.deg, u.deg))
+        center = skycoord_to_pixel(point, w, origin =1, mode = 'all') #This and previous line convert GAIA output source coordinates to pixel value
+        print(center)
+        if (center[0]  <= 0) or (center[0] >= 3054):
+            continue #makes sure sources are in frame
+        print('STEP3')
+        if (center[1] <= 0) or (center[1]  >= 2042): 
+            continue
+        print('STEP4')
+        x_cent = float(center[0])
+        y_cent = float(center[1])
+        x, y = centroid_sources(data, x_cent, y_cent, box_size = 30)
+        if math.isnan(x[0]) == True: #sometimes centroid sources returns unusable NAN values, this will reset them to GAIA centers if NAN.
+             x[0] = center[0]
+        if math.isnan(y[0]) == True:
+             y[0] = center[1]
+        bk_list = []
+        i = 0
+        while i != 100:
+            x_back = np.random.random(100) * 60 #use random smaples instead
+            y_back = np.random.random(100) * 60
+            back = np.stack((x_back, y_back), axis = -1)
+            avlist = []
+            for pnt in back:
+                if (x[0]+pnt[0]-30 <= 0) or (x[0]+pnt[0]-30 >= data.shape[1]): #making  sure to  exclude points outside the frame. Come up  with better way for   background sampling? Photutils????
+                    continue
+                if (y[0]+pnt[1]-30 <= 0) or (y[0]+pnt[1]-30 >= data.shape[0]):
+                    continue
+                avlist.append(data[int((y[0]-30)+pnt[1]), int((x[0]-30)+pnt[0])])
+            bk_list.append(np.sum(avlist)/(int(len(avlist))))
+            i +=1
+        AV = np.sum(bk_list)/(int(len(bk_list)))
+        C = 1
+        if data[int(y[0]), int(x[0])] > 10000:
+            C = Big_Fudge
+        if data[int(y[0]), int(x[0])] - AV  <  20:
+            C = small_fudge
+        print('Step5')
+        Rs = 4    
+        R1 = 0
+        while data[int(y[0]), int(x[0]+R1)] > AV:
+            R1 += .1
+            if (x[0]+R1) > data.shape[1]:
+                R1 =0
+                Rs -= 1
+                break
+            if R1 >= 17.5:
+                break
+        R2 = 0
+        while data[int(y[0]+R2), int(x[0])] > AV:
+            R2 += .1
+            if (y[0]+R2) > data.shape[0]:
+                R2 =0
+                Rs -= 1
+                break
+            if R2 >= 17.5:
+                break
+        R3 = 0
+        while data[int(y[0]), int(x[0]-R3)] > AV:
+            R3 += .1
+            if (x[0]-R3) < 0:
+                R3 =0
+                Rs -= 1
+                break
+            if R3 >= 17.5:
+                break
+        R4 = 0
+        while data[int(y[0]-R4), int(x[0])] > AV:
+            R4 += .1
+            if (y[0]-R4) > 0:
+                R4 =0
+                Rs -= 1
+                break
+            if R4 >= 17.5:
+                break
+        R = float(((R1+R2+R3+R4)/Rs)*C)
+        #print(R)
+        center_2 = PixCoord(x[0]+50, y[0]+50)
+        circle = CirclePixelRegion(center_2, R)
+        mask = circle.to_mask()
+        #data[mask.bbox.slices] *= 1 - mask.data
+        empty_mask[mask.bbox.slices] += mask
+    print('Step6')
+    dlt = np.arange(0, 50, 1)
+    np.delete(empty_mask, [i for i in dlt], axis= 0)
+    np.delete(empty_mask, [i+data.shape[0] for i in dlt], axis= 0)
+    np.delete(empty_mask, [i for i in dlt], axis= 1)
+    np.delete(empty_mask, [i+data.shape[1] for i in dlt], axis= 1)
+    #print(empty_mask.data)
+    print('Step7')
+    plt.imshow(empty_mask.data, norm=LogNorm(vmin=10, vmax=100))
+    lg = np.log(empty_mask.data+.001)
+    plt.imsave(fname='new_mask.png', arr=lg, cmap='viridis', dpi = 300) 
+    #plt.show()
+    #temp_im = fits.PrimaryHDU(data).writeto('/home/luca/Desktop/mask_test')
+    #return data
+
+#mask(img)