Upload New File

SimPlatform/Simulator.py

+# -*- coding: utf-8 -*-
+"""
+Created on Wed Aug  8 15:45:43 2018
+
+@author: Yi Zhang
+"""
+
+import numpy as np
+from scipy.signal import convolve2d
+from SimPlatform.ZoneTissue import ZoneTissue
+from SimPlatform.ZoneBubbles import ZoneBubbles
+from SimPlatform.Parameters import params_default
+from scipy.special import gamma
+
+"""Parameters for ultrasonic device"""
+        
+"""
+1.pixel: size of pixel of screen. unit=mm, pixel[0] corresponds to y dimension,
+         pixel[1] corresponds to x dimension.
+2.psf: the standard variation of the point spread function. psf[0] corresponds to
+        y dimension, psf[1] corresponds to x dimension.
+"""
+        
+class Simulator:
+    def __init__(self,params=params_default):
+        #global settings
+        self.pixel=params['pixel'] #size of pixel, unit: mm
+        self.shape=params['shape'] 
+        self.ampbubbles=params['ampbubbles']
+        self.amptissue=params['amptissue']
+        self.sigman=params['ampnoise']
+        self.sigman=self.sigman/np.sqrt(2)/gamma(1.5)
+        
+        #bubbles settings
+        self.zTissue=ZoneTissue(params)
+        self.zBubbles=ZoneBubbles(params)
+        #psf
+        s1,s2=params['psf']
+        s1=s1/self.pixel[0]
+        s2=s2/self.pixel[1]
+        fshape=[int(10*s1),int(10*s2)]
+        xc,yc=(fshape[1]-1)/2,(fshape[0]-1)/2
+        xv,yv=np.meshgrid(np.arange(fshape[1])-xc,np.arange(fshape[0])-yc)
+        self.filter=np.exp(-(xv**2/2/s2**2+yv**2/2/s1**2))
+        self.filter=self.filter.astype(np.float64)
+        
+    def generate(self,T=20):
+        Tissue=np.zeros(self.shape+tuple([T]),dtype=np.float64)
+        Bubbles=np.zeros(self.shape+tuple([T]),dtype=np.float64)
+        noi=np.zeros(self.shape+tuple([T]),dtype=np.float64)
+        #phi=np.random.uniform(0,2*np.pi)
+        #lmd=2*np.pi*15e6/34e4*self.pixel[0]
+        #_,yv=np.meshgrid(np.arange(self.shape[1]),np.arange(self.shape[0]))
+        #phase0=np.exp(1j*phi)*np.exp(1j*lmd*yv)
+        #phase=np.zeros(self.shape+tuple([T]),dtype=np.complex128)
+        
+        for t in range(T):
+            Bubbles[:,:,t]=np.real(self.zBubbles.image())
+            Tissue[:,:,t]=np.real(self.zTissue.image())
+            noi[:,:,t]=np.random.normal(0,self.sigman,self.shape) #+1j*np.random.normal(0,self.sigman,self.shape)
+            #phase[:,:,t]=phase0
+            
+            self.zBubbles.refresh()
+            self.zTissue.refresh()
+            
+        #control the amplitude of bubbles, tissue and noise
+        tmpb=Bubbles[Bubbles!=0]
+        if not (np.sum(np.abs(Bubbles))==0):
+            Bubbles=Bubbles*self.ampbubbles/np.mean(np.abs(tmpb))
+        AMtissue=np.random.uniform(0.6,1,[1])
+        Tissue=Tissue*self.amptissue/np.max(np.abs(Tissue))*2*AMtissue
+        
+        #point spread function
+        for i in range(T):
+            Bubbles[:,:,i]=self.psf(Bubbles[:,:,i])
+            Tissue[:,:,i]=self.psf(Tissue[:,:,i])
+            noi[:,:,i]=self.psf(noi[:,:,i])
+        Sum=Bubbles+Tissue+noi
+            
+        return Sum,Bubbles,Tissue
+    
+    def psf(self,A):
+        return convolve2d(A,self.filter,mode='same')
+        
+        
+        
+        
+        
\ No newline at end of file