class casoCalibExtr:
def __init__(self, fullData, intrCalibResults, case, stdPix, allDelta,
nCores, nTune, nTuneInter, tuneBool, tune_thr, tallyBool,
convChecksBool, rndSeedBool, scaNdim, scaAl, nDraws, nChains,
indexSave, pathFiles):
self.case = case
self.nCores = nCores
self.nTune = nTune
self.nTuneInter = nTuneInter
self.tuneBool = tuneBool
self.tune_thr = tune_thr
self.tallyBool = tallyBool
self.convChecksBool = convChecksBool
self.rndSeedBool = rndSeedBool
self.scaNdim = scaNdim
self.scaAl = scaAl
self.nDraws = nDraws
self.nChains = nChains
self.indexSave = indexSave
self.pathFiles = pathFiles
self.allDelta = allDelta
self.camera = fullData.Synt.Intr.camera
self.model = fullData.Synt.Intr.model
self.imgSize = fullData.Synt.Intr.s
Ns = [2, 3]
Xint = intrCalibResults['mean'][:3]
self.cameraMatrix, self.distCoeffs = bl.flat2int(Xint, Ns, self.model)
caseData = extractCaseData(exCase)
self.imagePoints = caseData[0]
self.xi, self.yi = self.imagePoints.T
self.objpoints = caseData[1]
self.rVecsT = caseData[2]
self.tVecsT = caseData[3]
self.xAllT = np.concatenate([self.rVecsT, self.tVecsT])
self.nPt = self.objpoints.shape[0]
self.nFree = 6
self.observedNormed = np.zeros((self.nPt * 2))
self.Crt = False # no RT error
self.Cf = np.zeros((4, 4))
self.Cf[2:, 2:] = intrCalibResults['cov'][:2, :2]
self.Ck = intrCalibResults['cov'][2, 2]
self.Cfk = np.zeros(4)
self.Cfk[2:] = intrCalibResults['cov'][:2, 2]
self.Ci = np.array([stdPix**2 * np.eye(2)] * self.nPt)
# tau de 1/5 pra la ventana movil
self.weiEsp = np.exp(- np.arange(nDraws) * 5 / nDraws)[::-1]
self.weiEsp /= np.sum(self.weiEsp)
def __init__(self, fullData, intrCalibResults, case, stdPix, allDelta,
nCores, nTune, nTuneInter, tuneBool, tune_thr, tallyBool,
convChecksBool, rndSeedBool, scaNdim, scaAl, nDraws, nChains,
indexSave, pathFiles):
self.case = case
self.nCores = nCores
self.nTune = nTune
self.nTuneInter = nTuneInter
self.tuneBool = tuneBool
self.tune_thr = tune_thr
self.tallyBool = tallyBool
self.convChecksBool = convChecksBool
self.rndSeedBool = rndSeedBool
self.scaNdim = scaNdim
self.scaAl = scaAl
self.nDraws = nDraws
self.nChains = nChains
self.indexSave = indexSave
self.pathFiles = pathFiles
self.allDelta = allDelta
self.camera = fullData.Synt.Intr.camera
self.model = fullData.Synt.Intr.model
self.imgSize = fullData.Synt.Intr.s
Ns = [2, 3]
Xint = intrCalibResults['mean'][:3]
self.cameraMatrix, self.distCoeffs = bl.flat2int(Xint, Ns, self.model)
caseData = extractCaseData(exCase)
self.imagePoints = caseData[0]
self.xi, self.yi = self.imagePoints.T
self.objpoints = caseData[1]
self.rVecsT = caseData[2]
self.tVecsT = caseData[3]
self.xAllT = np.concatenate([self.rVecsT, self.tVecsT])
self.nPt = self.objpoints.shape[0]
self.nFree = 6
self.observedNormed = np.zeros((self.nPt * 2))
self.Crt = False # no RT error
self.Cf = np.zeros((4, 4))
self.Cf[2:, 2:] = intrCalibResults['cov'][:2, :2]
self.Ck = intrCalibResults['cov'][2, 2]
self.Cfk = np.zeros(4)
self.Cfk[2:] = intrCalibResults['cov'][:2, 2]
self.Ci = np.array([stdPix**2 * np.eye(2)] * self.nPt)
# tau de 1/5 pra la ventana movil
self.weiEsp = np.exp(- np.arange(nDraws) * 5 / nDraws)[::-1]
self.weiEsp /= np.sum(self.weiEsp)
def perform(self, node, inputs_storage, output_storage):
# global projCount
# projCount += 1
# self.count += 1
# print('IDX %d projection %d, global %d' %
# (self.idx, self.count, projCount))
Xint, Xext = inputs_storage
# saco los parametros de flat para que los use la func de projection
# print(Xint)
cameraMatrix, distCoeffs = bl.flat2int(Xint, Ns, model)
# print(cameraMatrix, distCoeffs)
xy = np.zeros((n, m, 2))
cm = np.zeros((n, m, 2, 2))
for j in range(n):
rVec, tVec = bl.flat2ext(Xext[j])
xy[j, :, 0], xy[j, :, 1], cm[j] = cl.inverse(imagePoints.reshape(
(n, m, 2))[j],
rVec,
tVec,
cameraMatrix,
distCoeffs,
model,
Cccd=Ci[j])
# print(xy)
xy -= objpoints2D
S = np.linalg.inv(cm)
u, s, v = np.linalg.svd(S)
sdiag = np.zeros_like(u)
sdiag[:, :, [0, 1], [0, 1]] = np.sqrt(s)
A = (u.reshape((n, m, 2, 2, 1, 1)) * sdiag.reshape(
(n, m, 1, 2, 2, 1)) * v.transpose((0, 1, 3, 2)).reshape(
(n, m, 1, 1, 2, 2))).sum(axis=(3, 4))
xy = np.sum(xy.reshape((n, m, 2, 1)) * A, axis=3)
output_storage[0][0] = xy
def perform(self, node, inputs_storage, output_storage):
# global projCount
# projCount += 1
# self.count += 1
# print('IDX %d projection %d, global %d' %
# (self.idx, self.count, projCount))
Xint, Xext = inputs_storage
# saco los parametros de flat para que los use la func de projection
# print(Xint)
cameraMatrix, distCoeffs = bl.flat2int(Xint, Ns, model)
# print(cameraMatrix, distCoeffs)
xy = np.zeros((n, m, 2))
cm = np.zeros((n, m, 2, 2))
for j in range(n):
rVec, tVec = bl.flat2ext(Xext[j])
xy[j, :, 0], xy[j, :, 1], cm[j] = cl.inverse(
imagePoints.reshape((n, m, 2))[j], rVec, tVec,
cameraMatrix, distCoeffs, model, Cccd=Ci[j])
# print(xy)
xy -= objpoints2D
S = np.linalg.inv(cm)
u, s, v = np.linalg.svd(S)
sdiag = np.zeros_like(u)
sdiag[:,:,[0,1],[0,1]] = np.sqrt(s)
A = (u.reshape((n, m, 2, 2, 1, 1)) *
sdiag.reshape((n, m, 1, 2 ,2, 1)) *
v.transpose((0,1,3,2)).reshape((n, m, 1, 1, 2, 2))
).sum(axis=(3,4))
xy = np.sum(xy.reshape((n,m,2,1)) * A, axis=3)
output_storage[0][0] = xy
extrinsicFolder = "./resources/intrinsicCalib/" + camera + "/"
# load model specific data
#distCoeffs = np.load(distCoeffsFile)
#cameraMatrix = np.load(linearCoeffsFile)
# load intrinsic calib uncertainty
resultsML = np.load(intrinsicParamsFile).all() # load all objects
#Nmuestras = resultsML["Nsamples"]
Xint = resultsML['paramsMU']
covarDist = resultsML['paramsVAR']
#muCovar = resultsML['paramsMUvar']
#covarCovar = resultsML['paramsVARvar']
Ns = resultsML['Ns']
cameraMatrix, distCoeffs = bl.flat2int(Xint, Ns, model)
if model == modelos[3]:
Cf = np.zeros((4, 4))
Cf[2:, 2:] = covarDist[:2, :2]
Ck = covarDist[2, 2]
else:
Cf = covarDist[:4, :4]
Ck = covarDist[4:, 4:]
# Calibration points
dataPts = np.loadtxt(calibptsFile)
xIm = dataPts[:, :2] # PUNTOS DE CALIB EN PIXELES
n = 1
m = dataPts.shape[0]
fullDataFile = "./resources/fullDataIntrExtr.npy" dataFile = open(fullDataFile, "rb") fullData = pickle.load(dataFile) dataFile.close() # cam puede ser ['vca', 'vcaWide', 'ptz'] son los datos que se tienen camera = fullData.Synt.Intr.camera #modelos = ['poly', 'rational', 'fisheye', 'stereographic'] model = fullData.Synt.Intr.model Ns = [2, 3] nH = fullData.Synt.Extr.h.shape[0] nAng = fullData.Synt.Extr.ang.shape[0] fkV = np.concatenate([fullData.Synt.Intr.uv, [fullData.Synt.Intr.k]]) # # load model specific data cameraMatrix, distCoeffs = flat2int(fkV, Ns, model) # load data rVall = fullData.Synt.Extr.rVecs rVall = np.transpose([rVall] * nH, (1, 0, 2)).reshape((-1, 3)) tVall = fullData.Synt.Extr.tVecs.reshape((-1, 3)) rtVall = np.concatenate([rVall, tVall], axis=1) nPt = fullData.Synt.Extr.imgPt.shape[2] # cantidad de imagenes nIm = np.prod(fullData.Synt.Extr.imgPt.shape[:2]) # puntos por imagen # imagePoints = fullData.Synt.Extr.imgPt.reshape((nIm, nPt, 2)) chessboardModel = fullData.Synt.Ches.objPt imgSize = fullData.Synt.Intr.s # tomo las detecciones sin ruido
fullDataFile = "./resources/fullDataIntrExtr.npy" dataFile = open(fullDataFile, "rb") fullData = pickle.load(dataFile) dataFile.close() # cam puede ser ['vca', 'vcaWide', 'ptz'] son los datos que se tienen camera = fullData.Synt.Intr.camera #modelos = ['poly', 'rational', 'fisheye', 'stereographic'] model = fullData.Synt.Intr.model Ns = [2,3] nH = fullData.Synt.Extr.h.shape[0] nAng = fullData.Synt.Extr.ang.shape[0] fkV = np.concatenate([fullData.Synt.Intr.uv, [fullData.Synt.Intr.k]]) # # load model specific data cameraMatrix, distCoeffs = flat2int(fkV, Ns, model) # load data rVall = fullData.Synt.Extr.rVecs rVall = np.transpose([rVall] * nH, (1, 0, 2)).reshape((-1, 3)) tVall = fullData.Synt.Extr.tVecs.reshape((-1, 3)) rtVall = np.concatenate([rVall, tVall], axis=1) nPt = fullData.Synt.Extr.imgPt.shape[2] # cantidad de imagenes nIm = np.prod(fullData.Synt.Extr.imgPt.shape[:2]) # puntos por imagen # imagePoints = fullData.Synt.Extr.imgPt.reshape((nIm, nPt, 2)) chessboardModel = fullData.Synt.Ches.objPt imgSize = fullData.Synt.Intr.s # tomo las detecciones sin ruido
extrinsicFolder = "./resources/intrinsicCalib/" + camera + "/"
# load model specific data
#distCoeffs = np.load(distCoeffsFile)
#cameraMatrix = np.load(linearCoeffsFile)
# load intrinsic calib uncertainty
resultsML = np.load(intrinsicParamsFile).all() # load all objects
#Nmuestras = resultsML["Nsamples"]
Xint = resultsML['paramsMU']
covarDist = resultsML['paramsVAR']
#muCovar = resultsML['paramsMUvar']
#covarCovar = resultsML['paramsVARvar']
Ns = resultsML['Ns']
cameraMatrix, distCoeffs = bl.flat2int(Xint, Ns, model)
if model==modelos[3]:
Cf = np.zeros((4,4))
Cf[2:,2:] = covarDist[:2,:2]
Ck = covarDist[2,2]
else:
Cf = covarDist[:4,:4]
Ck = covarDist[4:,4:]
# Calibration points
dataPts = np.loadtxt(calibptsFile)
xIm = dataPts[:,:2] # PUNTOS DE CALIB EN PIXELES
n = 1