def Humoments(pca,cx,cy,p):
pca['raio'] = np.sqrt((pca['x'] - cx)**2 + (pca['y'] - cy)**2)
Rm = pca['raio'].mean()
SigR = pca['raio'].std()
Rm_re = pca['raio'].mean()/re
SigR_re = pca['raio'].std()/re
print('')
print('/'*70)
print('Hu moments - Populacao %d' %p)
print('Raio medio Rm: %5.3f +- %5.3f' %(Rm,SigR))
print('Raio madio ponderado Rm/Re: %5.3f +- %5.3f' %(Rm_re,SigR_re))
file = open('parametros_hu.txt', 'a')
'''
Momentos da imagem
Ver: http://pt.wikipedia.org/wiki/Momentos_Invariantes_de_uma_Imagem
'''
#nao centrais de 2 ordem
m11=(pca['x']*pca['y']).sum()
m20=(pca['x']*pca['x']).sum()
m02=(pca['y']*pca['y']).sum()
#centrais de 3 ordem
mu_11=((pca['x']-cx)*(pca['y']-cy)).sum()
mu_20=((pca['x']-cx)**2).sum()
mu_02=((pca['y']-cy)**2).sum()
mu_12=((pca['x']-cx)*((pca['y']-cy)**2)).sum()
mu_21=((pca['y']-cx)*((pca['x']-cx)**2)).sum()
mu_30=((pca['x']-cx)**3).sum()
mu_03=((pca['y']-cy)**3).sum()
#centrais de 4 ordem
mu_40=((pca['x']-cx)**4).sum()
mu_04=((pca['y']-cy)**4).sum()
#print('Momentos nao centrais m11, m20, m02: %d, %d, %d' %(m11, m20, m02))
#print('Momentos centrais mu_11, mu_20, mu_02: %d, %d, %d' %(mu_11, mu_20, mu_02))
#print('Momentos centrais mu_12, mu_21: %d, %d' %(mu_12, mu_21))
#Momentos invariantes por escala n_ij
n11=mu_11/(len(pca)**2)
n12=mu_12/(len(pca)**2.5)
n21=mu_21/(len(pca)**2.5)
n02=mu_02/(len(pca)**2)
n20=mu_20/(len(pca)**2)
n30=mu_30/(len(pca)**2.5)
n03=mu_03/(len(pca)**2.5)
#Momentos invariantes por translacao, escala e rotacao
#Invariantes de Hu
I1 = n02 + n20
I2 = ((n20-n02)**2) + 4*((n11)**2)
I3 = (n30 - 3*n12)**2 + (3*n21 - n03)**2
I4 = (n30 + 3*n12)**2 + (3*n21 + n03)**2
I5 = (n30 - 3*n12)*(n30 + n12)*(((n30 + n12)**2) - 3*((n21 + n03)**2)) + (3*n21 - n03)*(n12 + n03)*(3*((n30 + n12)**2) - ((n21 + n03)**2))
I6 = (n20 - n02)*((n30 + n12)**2 - (n21 + n03)**2) + 4*n11*(n30 + n12)*(n21 + n03)
I7 = (3*n21 - n03)*(n30 + n12)*((n30 + n12)**2 - 3*(n21 + n03)**2) - (n30 - 3*n12)*(n21 + n03)*(3*(n30 + n12)**2 - (n21 + n03)**2)
print('Hu Moments: %5.2e, %5.2e, %5.2e, %5.2e, %5.2e, %5.2e, %5.2e' %(I1, I2, I3, I4, I5, I6, I7))
#Parametros da Elipse
dd = (mu_20 + mu_02)
ee = (mu_20 - mu_02)*(mu_20 - mu_02) + 4*(mu_11)*(mu_11)
a = np.sqrt((2*(dd + np.sqrt(ee)))/len(pca))
b = np.sqrt((2*(dd - np.sqrt(ee)))/len(pca))
print('')
print('Parametros obtidos')
print('Semi-eixos da elipse (a,b): %5.3f, %5.3f' %(a,b))
#Razao dos semi-eixos
f = (a+b)/2
print('razao dos eixos (a+b/2): %5.3f' %f)
#Orientacao da Elipse
tetha = 0.5*np.arctan((2*mu_11)/(mu_20 - mu_02))
print('Orientacao da elipse, tetha: %5.3f' %tetha)
#Excentricidade
exc = 1 - (b/a)
print('Excentricidade da Elipse, e: %5.3f' %exc)
#FATOR DE ELONGACAO
flong = np.sqrt((mu_02/mu_20))
print('Fator de elongacao, flong: %5.3f' %flong)
file.write('%s %s %s %s %s %s %s %s %s %s %s %s %s %s %s %s\n' %(p, Rm_re, SigR_re, I1, I2, I3, I4, I5, I6, I7, a, b, f,
tetha, exc, flong))
'''
Simetria
'''
cte = cy - tetha*cx
sym1 = pca.ix[pca['y'] >= tetha*pca['x'] + cte]
sym2 = pca.ix[pca['y'] < tetha*pca['x'] + cte]
SYM = 1 - (math.fabs(len(sym1) - len(sym2))/(len(sym1) + len(sym2)))
print('')
print('Parametro de simetria: %5.4f' %SYM)
arr1=pca['x']
arr2=pca['y']
n=len(pca)
l=1
fof = fof_fortran.fof(arr1, arr2,l,n)
df_fof=pd.DataFrame(fof)
df_fof.columns = ['fof']
print(df_fof.describe())
return()
#if-then-else using numpys where()
df_rss2 = df_rss
df_rss2['logic'] = np.where(df_rss['flux_r'] > limiar, 1, 0)
np.savetxt('fof2.txt', df_rss2, fmt=formats2, delimiter='\t')
arr1 = df_seg['x']
arr2 = df_seg['y']
n = len(df_seg)
l = 1
arr = np.zeros(shape=len(df_seg))
arr3 = pd.DataFrame(arr)
a = fof_fortran.fof(arr1, arr2, arr3, l, n)
#a = fof_fortran.fof(arr1, arr2,l,n)
df_ff = pd.DataFrame(a)
df_ff.columns = ['fof']
compr = len(df_seg)
labels = []
#criando os rotulos iniciais
for row in range(0, compr):
labels.append(row)
df_seg['label'] = labels
df_ff['label'] = labels
cte = cy - tetha * cx
sym1 = data5.ix[data5['y'] >= tetha * data5['x'] + cte]
sym2 = data5.ix[data5['y'] < tetha * data5['x'] + cte]
print(len(sym1), len(sym2), len(data5), len(data4))
SYM = 1 - (math.fabs(len(sym1) - len(sym2)) / (len(sym1) + len(sym2)))
print('')
print('Parametro de simetria: %5.4f' % SYM)
arr1 = data5['x']
arr2 = data5['y']
n = len(data5)
l = 1
aaa = fof_fortran.fof(arr1, arr2, l, n)
df_ff = pd.DataFrame(aaa)
df_ff.columns = ['fof']
np.savetxt('bla.txt', df_ff)
print(df_ff.describe())
print('')
print('*' * 70)
fim = time.time()
time_proc = fim - ini
print('tempo de processamento: %fs' % time_proc)