python类convolve1d()的实例源码

def gaussian_smoothing(self, data=None, filter_len=None, filter_sigma=None):
    """ This method convolves the data with a gaussian
     the smoothed data is returned

    @param array data: raw data
    @param int filter_len: length of filter
    @param int filter_sigma: width of gaussian

    @return array: smoothed data

    """
    #Todo: Check for wrong data type
    if filter_len is None:
        if len(data) < 20.:
            filter_len = 5
        elif len(data) >= 100.:
            filter_len = 10
        else:
            filter_len = int(len(data) / 10.) + 1
    if filter_sigma is None:
        filter_sigma = filter_len

    gaus = gaussian(filter_len, filter_sigma)
    return filters.convolve1d(data, gaus / gaus.sum(), mode='mirror')

def iuwt(wave, convol2d =0):
    mode = 'nearest'

    lvl,n1,n2 = np.shape(wave)
    h = np.array([1./16, 1./4, 3./8, 1./4, 1./16])
    n = np.size(h)

    cJ = np.copy(wave[lvl-1,:,:])


    for i in np.linspace(1,lvl-1,lvl-1):

        newh = np.zeros((1,n+(n-1)*(2**(lvl-1-i)-1)))
        newh[0,np.int_(np.linspace(0,np.size(newh)-1,len(h)))] = h
        H = np.dot(newh.T,newh)

        ###### Line convolution
        if convol2d == 1:
            cnew = cp.convolve2d(cJ, H, mode='same', boundary='symm')
        else:
          cnew = sc.convolve1d(cJ,newh[0,:],axis = 0, mode = mode)
            ###### Column convolution
          cnew = sc.convolve1d(cnew,newh[0,:],axis = 1, mode = mode)

        cJ = cnew+wave[lvl-1-i,:,:]

    return np.reshape(cJ,(n1,n2))

def iuwt_1D(wave):
    """
    Inverse Starlet transform.
    INPUTS:
        wave: wavelet decomposition of an image.
    OUTPUTS:
        out: image reconstructed from wavelet coefficients
    OPTIONS:
        convol2d:  if set, a 2D version of the filter is used (slower, default is 0)

    """
    mode = 'nearest'

    lvl,n1= np.shape(wave)
    h = np.array([1./16, 1./4, 3./8, 1./4, 1./16])
    n = np.size(h)

    cJ = np.copy(wave[lvl-1,:])


    for i in np.linspace(1,lvl-1,lvl-1):

        newh = np.zeros((1,n+(n-1)*(2**(lvl-1-i)-1)))
        newh[0,np.int_(np.linspace(0,np.size(newh)-1,len(h)))] = h
        H = np.dot(newh.T,newh)

        ###### Line convolution

        cnew = sc.convolve1d(cJ,newh[0,:],axis = 0, mode = mode)
        cJ = cnew+wave[lvl-1-i,:]

    out = cJ
    return out

def estimate_poissonian(self, x_axis, data, params):
    """ Provide an estimator for initial values of a poissonian function.

    @param numpy.array x_axis: 1D axis values
    @param numpy.array data: 1D data, should have the same dimension as x_axis.
    @param lmfit.Parameters params: object includes parameter dictionary which
                                    can be set

    @return tuple (error, params):

    Explanation of the return parameter:
        int error: error code (0:OK, -1:error)
        Parameters object params: set parameters of initial values
    """

    error = self._check_1D_input(x_axis=x_axis, data=data, params=params)

    # a gaussian filter is appropriate due to the well approximation of poisson
    # distribution
    # gaus = gaussian(10,10)
    # data_smooth = filters.convolve1d(data, gaus/gaus.sum(), mode='mirror')
    data_smooth = self.gaussian_smoothing(data=data, filter_len=10,
                                          filter_sigma=10)

    # set parameters
    mu = x_axis[np.argmax(data_smooth)]
    params['mu'].value = mu
    params['amplitude'].value = data_smooth.max() / self.poisson(mu, mu)

    return error, params

def gaussianlinearoffset_testing_data():

    x = np.linspace(0, 5, 30)
    x_nice=np.linspace(0, 5, 101)

    mod_final,params = qudi_fitting.make_gaussianwithslope_model()

    data=np.loadtxt("./../1D_shllow.csv")
    data_noisy=data[:,1]
    data_fit=data[:,3]
    x=data[:,2]


    update=dict()
    update["slope"]={"min":-np.inf,"max":np.inf}
    update["offset"]={"min":-np.inf,"max":np.inf}
    update["sigma"]={"min":-np.inf,"max":np.inf}
    update["center"]={"min":-np.inf,"max":np.inf}
    update["amplitude"]={"min":-np.inf,"max":np.inf}
    result=qudi_fitting.make_gaussianwithslope_fit(x_axis=x, data=data_noisy, add_params=update)
#
##
#    gaus=gaussian(3,5)
#    qudi_fitting.data_smooth = filters.convolve1d(qudi_fitting.data_noisy, gaus/gaus.sum(),mode='mirror')

    plt.plot(x,data_noisy,label="data")
    plt.plot(x,data_fit,"k",label="old fit")
    plt.plot(x,result.init_fit,'-g',label='init')
    plt.plot(x,result.best_fit,'-r',label='fit')
    plt.legend()
    plt.show()
    print(result.fit_report())

def poissonian_testing():
    start=0
    stop=30
    mu=8
    num_points=1000
    x = np.array(np.linspace(start, stop, num_points))
#            x = np.array(x,dtype=np.int64)
    mod,params = qudi_fitting.make_poissonian_model()
    print('Parameters of the model',mod.param_names)

    p=Parameters()
    p.add('mu',value=mu)
    p.add('amplitude',value=200.)

    data_noisy=(mod.eval(x=x,params=p) *
                np.array((1+0.001*np.random.normal(size=x.shape) *
                p['amplitude'].value ) ) )

    print('all int',all(isinstance(item, (np.int32,int, np.int64)) for item in x))
    print('int',isinstance(x[1], int),float(x[1]).is_integer())
    print(type(x[1]))
    #make the filter an extra function shared and usable for other functions
    gaus=gaussian(10,10)
    data_smooth = filters.convolve1d(data_noisy, gaus/gaus.sum(),mode='mirror')


    result = qudi_fitting.make_poissonian_fit(x, data_noisy)
    print(result.fit_report())

    plt.figure()
    plt.plot(x, data_noisy, '-b', label='noisy data')
    plt.plot(x, data_smooth, '-g', label='smoothed data')
    plt.plot(x,result.init_fit,'-y', label='initial values')
    plt.plot(x,result.best_fit,'-r',linewidth=2.0, label='fit')
    plt.xlabel('counts')
    plt.ylabel('occurences')
    plt.legend(bbox_to_anchor=(0., 1.02, 1., .102), loc=3,
               ncol=2, mode="expand", borderaxespad=0.)
    plt.show()

def kernel(self, series, sigma=3):
        # fix the weight of data
        # http://www.nehalemlabs.net/prototype/blog/2014/04/12/
        #    how-to-fix-scipys-interpolating-spline-default-behavior/
        series = np.asarray(series)
        b = gaussian(25, sigma)
        averages = filters.convolve1d(series, b/b.sum())
        variances = filters.convolve1d(np.power(series-averages, 2), b/b.sum())
        variances[variances == 0] = 1
        return averages, variances

def temporal_feature_smoothing(video_features, kernel):
    #simple 1d convolution assuming that input is time x words x descriptors
    return convolve1d(video_features, weights = kernel, axis = 0)

def filterSimMat(simMat, filtLength, filtType, scaleFilterMethod='max1'):
    if filtType == 'hamming':
        filt = np.hamming(filtLength)
    elif filtType == 'flat':
        filt = np.ones(filtLength)
    else:
        raise RuntimeError("Unknown/unsupported filter type {}".format(filtType))
    if scaleFilterMethod == 'max1':
        filt /= np.max(filt)
    elif scaleFilterMethod == 'sum1':
        filt /= np.sum(filt)
    # print filt
    # filt = np.tile(filt, (simMat.shape[0], 1))
    # print filt.shape
    return filters.convolve1d(simMat, weights=filt, axis=1, mode='constant')

def _filterRows(X, filtLength):
    filt = np.hamming(filtLength)
    return filters.convolve1d(X, weights=filt, axis=1, mode='constant')

def filterRows(X, filtLength, filtType='hamming', scaleFilterMethod='max1'):
    if filtType == 'hamming':
        filt = np.hamming(filtLength)
    elif filtType == 'flat':
        filt = np.ones(filtLength)
    else:
        raise RuntimeError("Unknown/unsupported filter type {}".format(filtType))
    if scaleFilterMethod == 'max1':
        filt /= np.max(filt)
    elif scaleFilterMethod == 'sum1':
        filt /= np.sum(filt)

    return filters.convolve1d(X, weights=filt, axis=1, mode='constant')

def notSoRandomWalk(shape, std=1, trendFilterLength=32, lpfLength=16):
    """bandpass filter a random walk so that the low-frequency trend /
    drift is eliminated and the high-frequency noise is attenuated"""
    walk = randwalk(shape, std=std)
    filt = np.hamming(trendFilterLength)
    filt /= np.sum(filt)
    whichAxis = len(walk.shape) > 1 # 0 iff 1d, else 1
    # subtract baseline drift, roughly
    trend = filters.convolve1d(walk, weights=filt, axis=whichAxis, mode='reflect')
    walk -= trend
    # subtract noisey spikes
    walk = filters.convolve1d(walk, weights=np.hamming(lpfLength), axis=whichAxis, mode='reflect')
    return walk

def _filterRows(X, filtLength):
    filt = np.hamming(filtLength)
    return filters.convolve1d(X, weights=filt, axis=1, mode='constant')

def get_boundingbox(frame_set, use_region_of_interest=False):

    fstd = np.std(frame_set,axis=0)
    framesstd = np.mean(fstd)
    #th = framesstd  / 3
    th = framesstd
    #ones = np.ones(8)
    ones = np.ones(10)
    big_var = (fstd>th)

    if not use_region_of_interest or framesstd==0:
        # no bb, take full frame
        frameROIRes = np.zeros([20,50,50])
        for i in range(20):
            frameROIRes[i,:,:] = scipy.misc.imresize(frame_set[i,:,:], size=(50,50),interp='bilinear')
        #frameROIRes = np.reshape(frameROIRes, (1,frameROIRes.shape[0]*frameROIRes.shape[1]*frameROIRes.shape[2]))
        frameROIRes = frameROIRes.astype(np.float32)
        return frameROIRes  #, framesstd)

    big_var = big_var.astype(np.float32)
    big_var = filters.convolve1d(big_var, ones, axis=0)
    big_var = filters.convolve1d(big_var, ones, axis=1)

    th2 = 80
    i,j = np.nonzero(big_var>th2)

    if (i.size > 0):

        si = np.sort(i)
        sj = np.sort(j)


        ll = si.shape[0]
        th1 = int(round(ll*0.03))
        th2 = int(np.floor(ll*0.98))

        y1 = si[th1]
        y2 = si[th2]
        x1 = sj[th1]
        x2 = sj[th2]

        # cut image ROI
        if (((x2-x1)>0) and ((y2-y1)>0)):
            framesRoi = frame_set[:,y1:y2,x1:x2]
        else:
            framesRoi = frame_set[:,:,:]
    else:
        framesRoi = frame_set[:,:,:]

    # debug - show ROI
    #cv2.namedWindow('ROI', cv2.WINDOW_NORMAL)
    #bla= scipy.misc.imresize(framesRoi[19,:,:], size=(200,200),interp='bilinear')
    #cv2.imshow('ROI', bla)

    # resize to 50x50
    frameROIRes = np.zeros([20,50,50])
    for i in range(20):
        frameROIRes[i,:,:] = scipy.misc.imresize(framesRoi[i,:,:], size=(50,50),interp='bilinear')

    #frameROIRes = frameROIRes / 255  # TODO - does this really nessacarry?
    return (frameROIRes)

def get_boundingbox(self):

    fstd = numpy.std(self.frame_set,axis=0)
    framesstd = numpy.mean(fstd)
    th = framesstd
    ones = numpy.ones(10)
    big_var = (fstd>th)


    if (framesstd==0): # no bb, take full frame
        frameROIRes = numpy.zeros([20,50,50])
        for i in range(20):
        frameROIRes[i,:,:] = scipy.misc.imresize(self.frame_set[i,:,:], size=(50,50),interp='bilinear')

        frameROIRes = numpy.reshape(frameROIRes, (1,frameROIRes.shape[0]*frameROIRes.shape[1]*frameROIRes.shape[2]))
        frameROIRes = frameROIRes.astype(numpy.float32)

        return (frameROIRes)

    big_var = big_var.astype(numpy.float32)    
    big_var = filters.convolve1d(big_var, ones, axis=0)
    big_var = filters.convolve1d(big_var, ones, axis=1)

    th2 = 80
    i,j = numpy.nonzero(big_var>th2)

    if (i.size > 0):

        si = numpy.sort(i)
        sj = numpy.sort(j)


        ll = si.shape[0]
        th1 = round(ll*0.02)
        th2 = numpy.floor(ll*0.98)
        y1 = si[th1]
        y2 = si[th2]
        x1 = sj[th1]
        x2 = sj[th2]

        # cut image ROI
        if (((x2-x1)>0) and ((y2-y1)>0)):
        framesRoi = self.frame_set[:,y1:y2,x1:x2]
        else:       
        framesRoi = self.frame_set[:,:,:]
    else:
        framesRoi = self.frame_set[:,:,:]

    # resize to 50x50    
    frameROIRes = numpy.zeros([20,50,50])
    for i in range(20):
        frameROIRes[i,:,:] = scipy.misc.imresize(framesRoi[i,:,:], size=(50,50),interp='bilinear')

    riofstd = numpy.std(frameROIRes,axis=0)
    self.cur_std = numpy.mean(riofstd)

    frameROIRes = numpy.reshape(frameROIRes, (1,frameROIRes.shape[0]*frameROIRes.shape[1]*frameROIRes.shape[2]))
    frameROIRes = frameROIRes.astype(numpy.float32)

    return (frameROIRes)

def find_offset_parameter(self, x_values=None, data=None):
    """ This method convolves the data with a Lorentzian and the finds the
    offset which is supposed to be the most likely valy via a histogram.
    Additional the smoothed data is returned

    @param array x_values: x values
    @param array data: value of each data point corresponding to
                        x values

    @return int error: error code (0:OK, -1:error)
    @return float array data_smooth: smoothed data
    @return float offset: estimated offset


    """
    # lorentzian filter
    mod, params = self.make_lorentzian_model()

    # Todo: exclude filter in seperate method to be used in other methods

    if len(x_values) < 20.:
        len_x = 5
    elif len(x_values) >= 100.:
        len_x = 10
    else:
        len_x = int(len(x_values)/10.)+1

    lorentz = mod.eval(x=np.linspace(0, len_x, len_x), amplitude=1, offset=0.,
                       sigma=len_x/4., center=len_x/2.)
    data_smooth = filters.convolve1d(data, lorentz/lorentz.sum(),
                                     mode='constant', cval=data.max())

    # finding most frequent value which is supposed to be the offset
    hist = np.histogram(data_smooth, bins=10)
    offset = (hist[1][hist[0].argmax()]+hist[1][hist[0].argmax()+1])/2.

    return data_smooth, offset

############################################################################
#                                                                          #
#             Additional routines with gaussian-like filter              #
#                                                                          #
############################################################################