Loading csst_mci_sim/support/cosmicrays.py 0 → 100644 +271 −0 Original line number Diff line number Diff line """ Cosmic Rays =========== This simple class can be used to include cosmic ray events to an image. By default the cosmic ray events are drawn from distributions describing the length and energy of the events. Such distributions can be generated for example using Stardust code (http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=04636917). The energy of the cosmic ray events can also be set to constant for testing purposes. The class can be used to draw a single cosmic ray event or up to a covering fraction. :requires: NumPy :requires: SciPy :version: 0.2 """ import numpy as np from scipy.interpolate import InterpolatedUnivariateSpline class cosmicrays(): """ Cosmic ray generation class. Can either draw events from distributions or set the energy of the events to a constant. :param log: logger instance :param image: image to which cosmic rays are added to (a copy is made not to change the original numpy array) :param crInfo: column information (cosmic ray file) :param information: cosmic ray track information (file containing track length and energy information) and exposure time. """ def __init__(self, log, image, exptime, crInfo=None, information=None): """ Cosmic ray generation class. Can either draw events from distributions or set the energy of the events to a constant. :param log: logger instance :param image: image to which cosmic rays are added to (a copy is made not to change the original numpy array) :param crInfo: column information (cosmic ray file) :param information: cosmic ray track information (file containing track length and energy information) and exposure time. """ self.exptime = exptime self.log = log #image and size self.image = image.copy() self.ysize, self.xsize = self.image.shape if crInfo is not None: self.cr = crInfo else: self._readCosmicrayInformation() ################################################## def _cosmicRayIntercepts(self, lum, x0, y0, l, phi): """ Derive cosmic ray streak intercept points. :param lum: luminosities of the cosmic ray tracks :param x0: central positions of the cosmic ray tracks in x-direction :param y0: central positions of the cosmic ray tracks in y-direction :param l: lengths of the cosmic ray tracks :param phi: orientation angles of the cosmic ray tracks :return: cosmic ray map (image) :rtype: nd-array """ # create empty array crImage = np.zeros((self.ysize, self.xsize), dtype=np.float64) # x and y shifts dx = l * np.cos(phi) / 2. dy = l * np.sin(phi) / 2. mskdx = np.abs(dx) < 1e-8 mskdy = np.abs(dy) < 1e-8 dx[mskdx] = 0. dy[mskdy] = 0. #pixels in x-direction ilo = np.round(x0.copy() - dx) msk = ilo < 0. ilo[msk] = 0 ilo = ilo.astype(int) ihi = 1 + np.round(x0.copy() + dx) msk = ihi > self.xsize ihi[msk] = self.xsize ihi = ihi.astype(int) #pixels in y-directions jlo = np.round(y0.copy() - dy) msk = jlo < 0. jlo[msk] = 0 jlo = jlo.astype(int) jhi = 1 + np.round(y0.copy() + dy) msk = jhi > self.ysize jhi[msk] = self.ysize jhi = jhi.astype(int) # loop over the individual events for i, luminosity in enumerate(lum): n = 0 # count the intercepts u = [] x = [] y = [] # Compute X intercepts on the pixel grid if ilo[i] < ihi[i]: for xcoord in range(ilo[i], ihi[i]): ok = (xcoord - x0[i]) / dx[i] if np.abs(ok) <= 0.5: n += 1 u.append(ok) x.append(xcoord) y.append(y0[i] + ok * dy[i]) else: for xcoord in range(ihi[i], ilo[i]): ok = (xcoord - x0[i]) / dx[i] if np.abs(ok) <= 0.5: n += 1 u.append(ok) x.append(xcoord) y.append(y0[i] + ok * dy[i]) # Compute Y intercepts on the pixel grid if jlo[i] < jhi[i]: for ycoord in range(jlo[i], jhi[i]): ok = (ycoord - y0[i]) / dy[i] if np.abs(ok) <= 0.5: n += 1 u.append(ok) x.append(x0[i] + ok * dx[i]) y.append(ycoord) else: for ycoord in range(jhi[i], jlo[i]): ok = (ycoord - y0[i]) / dy[i] if np.abs(ok) <= 0.5: n += 1 u.append(ok) x.append(x0[i] + ok * dx[i]) y.append(ycoord) # check if no intercepts were found if n < 1: xc = int(np.floor(x0[i])) yc = int(np.floor(y0[i])) crImage[yc, xc] += luminosity # Find the arguments that sort the intersections along the track u = np.asarray(u) x = np.asarray(x) y = np.asarray(y) args = np.argsort(u) u = u[args] x = x[args] y = y[args] # Decide which cell each interval traverses, and the path length for i in range(1, n - 1): w = u[i + 1] - u[i] cx = int(1 + np.floor((x[i + 1] + x[i]) / 2.)) cy = int(1 + np.floor((y[i + 1] + y[i]) / 2.)) if 0 <= cx < self.xsize and 0 <= cy < self.ysize: crImage[cy, cx] += (w * luminosity) return crImage ############################################################################ def _drawEventsToCoveringFactor(self, coveringFraction=3.0, limit=1000, verbose=False): """ Generate cosmic ray events up to a covering fraction and include it to a cosmic ray map (self.cosmicrayMap). :param coveringFraction: covering fraction of cosmic rya events in per cent of total number of pixels :type coveringFraction: float :param limit: limiting energy for the cosmic ray event [None = draw from distribution] :type limit: None or float :param verbose: print out information to stdout :type verbose: bool :return: None """ self.cosmicrayMap = np.zeros((self.ysize, self.xsize)) # how many events to draw at once, too large number leads to exceeding the covering fraction ####cr_n = int(295 * self.exptime / 565. * coveringFraction / 1.4) cr_n = int(5000 * self.exptime / 565. * coveringFraction) covering = 0.0 while covering < coveringFraction: # pseudo-random numbers taken from a uniform distribution between 0 and 1 np.random.seed() luck = np.random.rand(cr_n) # draw the length of the tracks ius = InterpolatedUnivariateSpline( self.cr['cr_cdf'], self.cr['cr_u']) self.cr['cr_l'] = ius(luck) if limit is None: ius = InterpolatedUnivariateSpline( self.cr['cr_cde'], self.cr['cr_v']) self.cr['cr_e'] = ius(luck) else: # set the energy directly to the limit self.cr['cr_e'] = np.asarray([limit,]) # Choose the properties such as positions and an angle from a random Uniform dist np.random.seed() cr_x = self.xsize * np.random.rand(int(np.floor(cr_n))) np.random.seed() cr_y = self.ysize * np.random.rand(int(np.floor(cr_n))) np.random.seed() cr_phi = np.pi * np.random.rand(int(np.floor(cr_n))) # find the intercepts self.cosmicrayMap += self._cosmicRayIntercepts( self.cr['cr_e'], cr_x, cr_y, self.cr['cr_l'], cr_phi) # count the covering factor area_cr = np.count_nonzero(self.cosmicrayMap) covering = 100.*area_cr / (self.xsize*self.ysize) text = 'The cosmic ray covering factor is %i pixels i.e. %.3f per cent' % ( area_cr, covering) self.log.info(text) # 33 def addUpToFraction(self, coveringFraction, limit=None, verbose=False): """ Add cosmic ray events up to the covering Fraction. :param coveringFraction: covering fraction of cosmic rya events in per cent of total number of pixels :type coveringFraction: float :param limit: limiting energy for the cosmic ray event [None = draw from distribution] :type limit: None or float :param verbose: print out information to stdout :type verbose: bool :return: image with cosmic rays :rtype: ndarray """ self._drawEventsToCoveringFactor( coveringFraction, limit=limit, verbose=verbose) # paste cosmic rays self.image += self.cosmicrayMap return self.image Loading
csst_mci_sim/support/cosmicrays.py 0 → 100644 +271 −0 Original line number Diff line number Diff line """ Cosmic Rays =========== This simple class can be used to include cosmic ray events to an image. By default the cosmic ray events are drawn from distributions describing the length and energy of the events. Such distributions can be generated for example using Stardust code (http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=04636917). The energy of the cosmic ray events can also be set to constant for testing purposes. The class can be used to draw a single cosmic ray event or up to a covering fraction. :requires: NumPy :requires: SciPy :version: 0.2 """ import numpy as np from scipy.interpolate import InterpolatedUnivariateSpline class cosmicrays(): """ Cosmic ray generation class. Can either draw events from distributions or set the energy of the events to a constant. :param log: logger instance :param image: image to which cosmic rays are added to (a copy is made not to change the original numpy array) :param crInfo: column information (cosmic ray file) :param information: cosmic ray track information (file containing track length and energy information) and exposure time. """ def __init__(self, log, image, exptime, crInfo=None, information=None): """ Cosmic ray generation class. Can either draw events from distributions or set the energy of the events to a constant. :param log: logger instance :param image: image to which cosmic rays are added to (a copy is made not to change the original numpy array) :param crInfo: column information (cosmic ray file) :param information: cosmic ray track information (file containing track length and energy information) and exposure time. """ self.exptime = exptime self.log = log #image and size self.image = image.copy() self.ysize, self.xsize = self.image.shape if crInfo is not None: self.cr = crInfo else: self._readCosmicrayInformation() ################################################## def _cosmicRayIntercepts(self, lum, x0, y0, l, phi): """ Derive cosmic ray streak intercept points. :param lum: luminosities of the cosmic ray tracks :param x0: central positions of the cosmic ray tracks in x-direction :param y0: central positions of the cosmic ray tracks in y-direction :param l: lengths of the cosmic ray tracks :param phi: orientation angles of the cosmic ray tracks :return: cosmic ray map (image) :rtype: nd-array """ # create empty array crImage = np.zeros((self.ysize, self.xsize), dtype=np.float64) # x and y shifts dx = l * np.cos(phi) / 2. dy = l * np.sin(phi) / 2. mskdx = np.abs(dx) < 1e-8 mskdy = np.abs(dy) < 1e-8 dx[mskdx] = 0. dy[mskdy] = 0. #pixels in x-direction ilo = np.round(x0.copy() - dx) msk = ilo < 0. ilo[msk] = 0 ilo = ilo.astype(int) ihi = 1 + np.round(x0.copy() + dx) msk = ihi > self.xsize ihi[msk] = self.xsize ihi = ihi.astype(int) #pixels in y-directions jlo = np.round(y0.copy() - dy) msk = jlo < 0. jlo[msk] = 0 jlo = jlo.astype(int) jhi = 1 + np.round(y0.copy() + dy) msk = jhi > self.ysize jhi[msk] = self.ysize jhi = jhi.astype(int) # loop over the individual events for i, luminosity in enumerate(lum): n = 0 # count the intercepts u = [] x = [] y = [] # Compute X intercepts on the pixel grid if ilo[i] < ihi[i]: for xcoord in range(ilo[i], ihi[i]): ok = (xcoord - x0[i]) / dx[i] if np.abs(ok) <= 0.5: n += 1 u.append(ok) x.append(xcoord) y.append(y0[i] + ok * dy[i]) else: for xcoord in range(ihi[i], ilo[i]): ok = (xcoord - x0[i]) / dx[i] if np.abs(ok) <= 0.5: n += 1 u.append(ok) x.append(xcoord) y.append(y0[i] + ok * dy[i]) # Compute Y intercepts on the pixel grid if jlo[i] < jhi[i]: for ycoord in range(jlo[i], jhi[i]): ok = (ycoord - y0[i]) / dy[i] if np.abs(ok) <= 0.5: n += 1 u.append(ok) x.append(x0[i] + ok * dx[i]) y.append(ycoord) else: for ycoord in range(jhi[i], jlo[i]): ok = (ycoord - y0[i]) / dy[i] if np.abs(ok) <= 0.5: n += 1 u.append(ok) x.append(x0[i] + ok * dx[i]) y.append(ycoord) # check if no intercepts were found if n < 1: xc = int(np.floor(x0[i])) yc = int(np.floor(y0[i])) crImage[yc, xc] += luminosity # Find the arguments that sort the intersections along the track u = np.asarray(u) x = np.asarray(x) y = np.asarray(y) args = np.argsort(u) u = u[args] x = x[args] y = y[args] # Decide which cell each interval traverses, and the path length for i in range(1, n - 1): w = u[i + 1] - u[i] cx = int(1 + np.floor((x[i + 1] + x[i]) / 2.)) cy = int(1 + np.floor((y[i + 1] + y[i]) / 2.)) if 0 <= cx < self.xsize and 0 <= cy < self.ysize: crImage[cy, cx] += (w * luminosity) return crImage ############################################################################ def _drawEventsToCoveringFactor(self, coveringFraction=3.0, limit=1000, verbose=False): """ Generate cosmic ray events up to a covering fraction and include it to a cosmic ray map (self.cosmicrayMap). :param coveringFraction: covering fraction of cosmic rya events in per cent of total number of pixels :type coveringFraction: float :param limit: limiting energy for the cosmic ray event [None = draw from distribution] :type limit: None or float :param verbose: print out information to stdout :type verbose: bool :return: None """ self.cosmicrayMap = np.zeros((self.ysize, self.xsize)) # how many events to draw at once, too large number leads to exceeding the covering fraction ####cr_n = int(295 * self.exptime / 565. * coveringFraction / 1.4) cr_n = int(5000 * self.exptime / 565. * coveringFraction) covering = 0.0 while covering < coveringFraction: # pseudo-random numbers taken from a uniform distribution between 0 and 1 np.random.seed() luck = np.random.rand(cr_n) # draw the length of the tracks ius = InterpolatedUnivariateSpline( self.cr['cr_cdf'], self.cr['cr_u']) self.cr['cr_l'] = ius(luck) if limit is None: ius = InterpolatedUnivariateSpline( self.cr['cr_cde'], self.cr['cr_v']) self.cr['cr_e'] = ius(luck) else: # set the energy directly to the limit self.cr['cr_e'] = np.asarray([limit,]) # Choose the properties such as positions and an angle from a random Uniform dist np.random.seed() cr_x = self.xsize * np.random.rand(int(np.floor(cr_n))) np.random.seed() cr_y = self.ysize * np.random.rand(int(np.floor(cr_n))) np.random.seed() cr_phi = np.pi * np.random.rand(int(np.floor(cr_n))) # find the intercepts self.cosmicrayMap += self._cosmicRayIntercepts( self.cr['cr_e'], cr_x, cr_y, self.cr['cr_l'], cr_phi) # count the covering factor area_cr = np.count_nonzero(self.cosmicrayMap) covering = 100.*area_cr / (self.xsize*self.ysize) text = 'The cosmic ray covering factor is %i pixels i.e. %.3f per cent' % ( area_cr, covering) self.log.info(text) # 33 def addUpToFraction(self, coveringFraction, limit=None, verbose=False): """ Add cosmic ray events up to the covering Fraction. :param coveringFraction: covering fraction of cosmic rya events in per cent of total number of pixels :type coveringFraction: float :param limit: limiting energy for the cosmic ray event [None = draw from distribution] :type limit: None or float :param verbose: print out information to stdout :type verbose: bool :return: image with cosmic rays :rtype: ndarray """ self._drawEventsToCoveringFactor( coveringFraction, limit=limit, verbose=verbose) # paste cosmic rays self.image += self.cosmicrayMap return self.image
