zeissimporter.py 11.1 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
# -*- coding: utf-8 -*-
"""
GEPARD - Gepard-Enabled PARticle Detection
Copyright (C) 2018  Lars Bittrich and Josef Brandt, Leibniz-Institut für 
Polymerforschung Dresden e. V. <bittrich-lars@ipfdd.de>    

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, either version 3 of the License, or
(at your option) any later version.

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, see COPYING.  
If not, see <https://www.gnu.org/licenses/>.
"""
import os
from PyQt5 import QtCore, QtWidgets
23 24 25 26 27
from .zeissxml import ZeissHandler, make_parser
from .opticalscan import PointCoordinates
from .helperfunctions import cv2imread_fix, cv2imwrite_fix
from .ramancom.ramancontrol import defaultPath
from .dataset import DataSet
28
from scipy.optimize import least_squares
29
from itertools import permutations
30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231
import cv2
import numpy as np

class ZeissImporter(QtWidgets.QDialog):
    def __init__(self, fname, ramanctrl, parent=None):
        super().__init__(parent)
        
        if not ramanctrl.connect() or not self.readImportData(fname):
            msg = QtWidgets.QMessageBox()
            msg.setText('Connection failed! Please enable remote control.')
            msg.exec()
            self.validimport = False
            return
        else:
            self.validimport = True
            
        self.ramanctrl = ramanctrl
        
        vbox = QtWidgets.QVBoxLayout()
        pointgroup = QtWidgets.QGroupBox('Marker coordinates at Raman spot [µm]', self)
        self.points = PointCoordinates(len(self.markers), self.ramanctrl, self,
                                       names = [m.name for m in self.markers])
        self.points.pimageOnly.setVisible(False)
        pointgroup.setLayout(self.points)
        self.points.readPoint.connect(self.takePoint)
        
        self.pconvert = QtWidgets.QPushButton('Convert', self)
        self.pexit = QtWidgets.QPushButton('Cancel', self)
        self.pconvert.released.connect(self.convert)
        self.pexit.released.connect(self.reject)
        self.pconvert.setEnabled(False)
        
        btnLayout = QtWidgets.QHBoxLayout()
        btnLayout.addStretch()
        btnLayout.addWidget(self.pconvert)
        btnLayout.addWidget(self.pexit)
        
        label = QtWidgets.QLabel("Z-Image blur radius", self)
        self.blurspinbox = QtWidgets.QSpinBox(self)
        self.blurspinbox.setMinimum(3)
        self.blurspinbox.setMaximum(99)
        self.blurspinbox.setSingleStep(2)
        self.blurspinbox.setValue(5)
        blurlayout = QtWidgets.QHBoxLayout()
        blurlayout.addWidget(label)
        blurlayout.addWidget(self.blurspinbox)
        blurlayout.addStretch()
        
        vbox.addWidget(pointgroup)
        vbox.addLayout(blurlayout)
        vbox.addLayout(btnLayout)
        self.setLayout(vbox)
        
    def readImportData(self, fname):
        path = os.path.split(fname)[0]
        self.zmapimgname = os.path.join(path, '3D.tif')
        self.edfimgname = os.path.join(path, 'EDF.tif')
        xmlname = os.path.join(path, '3D.tif_metadata.xml')

        errmsges = []
        if not os.path.exists(self.zmapimgname):
            errmsges.append('Depth map image not found: 3D.tif')
        if not os.path.exists(self.edfimgname):
            errmsges.append('EDF image not found: EDF.tif')
        if not os.path.exists(xmlname):
            errmsges.append('XML metadata not found: 3D.tif_metadata.xml')
        else:
            parser = make_parser()
            z = ZeissHandler()
            parser.setContentHandler(z)
            parser.parse(xmlname)
            
            if len(z.markers)<3:
                errmsges.append('Fewer than 3 markers found to adjust coordinates!')
            if None in [z.region.centerx, z.region.centery, 
                        z.region.width, z.region.height]:
                errmsges.append('Image dimensions incomplete or missing!')
            if None in [z.zrange.z0, z.zrange.zn, z.zrange.dz]:
                errmsges.append('ZStack information missing or incomplete!')
    
        if len(errmsges)>0:
            QtWidgets.QMessageBox.error(self, 'Error!',
                                '\n'.join(errmsges),
                                QtWidgets.QMessageBox.Ok, 
                                QtWidgets.QMessageBox.Ok)
            return False
        
        self.region = z.region
        self.zrange = z.zrange
        self.markers = z.markers
        return True
        
    @QtCore.pyqtSlot(float, float, float) 
    def takePoint(self, x, y, z):
        points = self.points.getPoints()
        if len(points)>=3:
            self.pconvert.setEnabled(True)
     
    @QtCore.pyqtSlot()
    def convert(self):
        fname = QtWidgets.QFileDialog.getSaveFileName(self,
                           'Create New GEPARD Project', defaultPath, '*.pkl')[0]
        if fname=='':
            return
        dataset = DataSet(fname, newProject=True)
        T, pc, zpc = self.getTransform()
        imgshape, warp_mat = self.convertZimg(dataset, T, pc, zpc)
        self.convertImage(dataset, warp_mat)
        dataset.save()
        self.gepardname = dataset.fname
        self.accept()
        
    def convertImage(self, dataset, warp_mat):
        img = cv2imread_fix(self.edfimgname)
        img = cv2.warpAffine(img, warp_mat, img.shape[:2][::-1])
        cv2imwrite_fix(dataset.getImageName(), img)
        
    
    def convertZimg(self, dataset, T, pc, zpc):
        N = int(round((self.zrange.zn-self.zrange.z0)/self.zrange.dz))
        dataset.zpositions = np.linspace(self.zrange.z0, 
                                         self.zrange.zn, N)-zpc[2]+pc[2]
        zimg = cv2imread_fix(self.zmapimgname, cv2.IMREAD_GRAYSCALE)
        zmdist = zimg.mean()
        zm = zmdist/255.*(self.zrange.zn-self.zrange.z0) + self.zrange.z0
        
        radius = self.blurspinbox.value()
        blur = cv2.GaussianBlur(zimg, (radius, radius), 0)

        pshift = self.ramanctrl.getRamanPositionShift()
        dataset.pshift = pshift
        pixelscale = self.region.width/zimg.shape[1]
        
        # use input image as single image aquired in one shot
        dataset.imagedim_df = (self.region.width, self.region.height, 0.0)
        dataset.pixelscale_df = pixelscale
        
        dataset.imagedim_bf = (self.region.width, self.region.height, 0.0)
        dataset.pixelscale_bf = pixelscale
        
        # set image center as reference point in data set (transform from Zeiss)
        p0 = np.dot((np.array([self.region.centerx, 
                               self.region.centery,zm])-zpc),T)[:2] + pc[:2]
        dataset.readin = False
        dataset.lastpos = p0
        dataset.maxdim = p0 + p0
        
        # pixel triangle for coordinate warping transformation
        srcTri = np.array( [[0, 0], [zimg.shape[1] - 1, 0], 
                            [0, zimg.shape[0] - 1]] ).astype(np.float32)
        # upper left point (0,0) in Zeiss coordinates:
        z0 = np.array([self.region.centerx - self.region.width/2, 
                       self.region.centery + self.region.height/2])
        # transform pixel data to Zeiss coordinates
        dstTri = np.array([[p[0]*pixelscale + z0[0], 
                            z0[1] - p[1]*pixelscale, zm] for p in srcTri]).astype(np.double)-zpc
        
        # transform to Raman coordinates
        dstTri = np.dot(dstTri,T) + pc[np.newaxis,:]
        
        # tilt blur image based on transformend z and adapt zpositions
        x = np.linspace(0,1,blur.shape[1])
        y = np.linspace(0,1,blur.shape[0])
        x, y = np.meshgrid(x,y)
        zmap = x*(dstTri[1,2]-dstTri[0,2]) + y*(dstTri[2,2]-dstTri[0,2]) + \
               (zimg * ((self.zrange.zn-self.zrange.z0)/255.) - \
                zmdist*((self.zrange.zn-self.zrange.z0)/255.))
        zmin, zmax = zmap.min(), zmap.max()
        dataset.zpositions = np.array([zmap.min(), zmap.max()])
        blur = (zmap-zmin)*(255./(zmax-zmin))
        blur[blur>255.] = 255.
        blur = np.uint8(blur)
        # transform triangle back to pixel
        dstTri = np.array([dataset.mapToPixel(p[:2]) for p in dstTri]).astype(np.float32)
        
        warp_mat = cv2.getAffineTransform(srcTri, dstTri)

        blur = cv2.warpAffine(blur, warp_mat, zimg.shape[::-1])
        zimgname = dataset.getZvalImageName()
        cv2imwrite_fix(zimgname, blur)
        return zimg.shape, warp_mat
    
    def getTransform(self):
        points = self.points.getPoints()
        pshift = self.ramanctrl.getRamanPositionShift()
        points[:,0] -= pshift[0]
        points[:,1] -= pshift[1]
        zpoints = np.array([m.getPos() for m in self.markers], dtype=np.double)
        pc = points.mean(axis=0)
        zpc = zpoints.mean(axis=0)
        
        points -= pc[np.newaxis,:]
        zpoints -= zpc[np.newaxis,:]
        
        def getRotMat(angles):
            c1, s1 = np.cos(angles[0]), np.sin(angles[0])
            c2, s2 = np.cos(angles[1]), np.sin(angles[1])
            c3, s3 = np.cos(angles[2]), np.sin(angles[2])
            return np.mat([[c1*c3-s1*c2*s3, -c1*s3-s1*c2*c3, s1*s2],
                           [s1*c3+c1*c2*s3, -s1*s3+c1*c2*c3, -c1*s2],
                           [s1*s3, s2*c3, c2]])
        
232 233 234 235 236 237 238 239 240 241 242
        # find the transformation matrix with best fit for small angles in 
        # [-45°,45°] for all permutation of markers
        permbest = None
        pointsbest = None
        for perm in permutations(range(points.shape[0])):
            ppoints = points[perm,:]
            
            def err(angles_shift):
                T = getRotMat(angles_shift[:3]).T.A
                return (np.dot(zpoints, T) - angles_shift[np.newaxis,3:] \
                        - ppoints).ravel()
243

244 245 246 247 248 249 250 251 252 253 254 255 256 257
            angle = np.zeros(3)
            opt = least_squares(err, np.concatenate((angle, np.zeros(3))), 
                                bounds=(np.array([-np.pi/4]*3+[-np.inf]*3),
                                        np.array([np.pi/4]*3+[np.inf]*3)),
                                method='dogbox')
            
            if permbest is None or \
                permbest.cost>opt.cost:
                print("Current best permutation:", perm, flush=True)
                permbest = opt
                pointsbest = ppoints
        
        optangles = permbest.x[:3]
        shift = permbest.x[3:]
258
        T = getRotMat(optangles).T.A
259
        e = (np.dot(zpoints, T)-shift[np.newaxis,:]-pointsbest)
260 261 262 263 264 265 266 267 268 269 270
        print("Transformation angles:", optangles, flush=True)
        print("Transformation shift:", shift, flush=True)
        print("Transformation err:", e, flush=True)
        d = np.linalg.norm(e, axis=1)
        if np.any(d>1.):
            QtWidgets.QMessageBox.warning(self, 'Warning!',
                                f'Transformation residuals are large:{d}',
                                QtWidgets.QMessageBox.Ok, 
                                QtWidgets.QMessageBox.Ok)
        return T, pc-shift, zpc