def semantic3d_sample_block(beg, bi, ri, rm, fs, fn, min_p, bsize, ds_stride):
# from io_util import get_semantic3d_class_colors
# colors=get_semantic3d_class_colors()
lbls, pts = [], []
pn=0
for i in xrange(fn):
points, labels = read_room_pkl('data/Semantic3D.Net/pkl/train/' + fs + '_{}.pkl'.format(i))
idxs = libPointUtil.gridDownsampleGPU(points, ds_stride, False)
points = points[idxs]
labels = labels[idxs]
points[:, :3] = np.dot(points[:, :3], rm)
points[:, :3] -= np.expand_dims(min_p, axis=0)
x_cond = (points[:, 0] >= beg[0]) & (points[:, 0] < beg[0] + bsize)
y_cond = (points[:, 1] >= beg[1]) & (points[:, 1] < beg[1] + bsize)
cond = x_cond & y_cond
pn+=np.sum(cond)
pts.append(points[cond])
lbls.append(labels[cond])
if pn>1024:
pts = np.concatenate(pts, axis=0)
lbls = np.concatenate(lbls, axis=0)
print 'block {} pn {}'.format(bi,pn)
save_room_pkl('data/Semantic3D.Net/block/train/' + fs + '_{}_{}.pkl'.format(bi,ri), pts, lbls)
def prepare_semantic3d():
with open('cached/semantic3d_stems.txt','r') as f:
fns=f.readlines()
fns=[fn.strip('\n') for fn in fns]
for fn in fns:
labels=read_semantic3d_label_file('/home/pal/data/Semantic3D.Net/raw/train/'+fn+'.labels')
points=read_semantic3d_points_file('/home/pal/data/Semantic3D.Net/raw/train/'+fn+'.txt')
save_room_pkl('/home/pal/data/Semantic3D.Net/pkl/train/'+fn+'.pkl',points,labels)
print 'done'
def prepare_semantic3d_partition():
with open('cached/semantic3d_stems.txt','r') as f:
fns=f.readlines()
fns=[fn.strip('\n').split(' ')[0] for fn in fns]
for fn in fns[:7]:
lf=open('data/Semantic3D.Net/raw/train/' + fn + '.labels')
pf=open('data/Semantic3D.Net/raw/train/' + fn + '.txt')
count=0
while True:
line=lf.readline()
if not line:
break
count+=1
lf.seek(0)
points,labels=[],[]
pi,part_id=0,0
while True:
line=pf.readline()
if not line:
break
line=line.strip('\n')
sublines=line.split(' ')
x=float(sublines[0])
y=float(sublines[1])
z=float(sublines[2])
intensity=float(sublines[3])
r=float(sublines[4])
g=float(sublines[5])
b=float(sublines[6])
points.append(np.asarray([x,y,z,r,g,b,intensity],dtype=np.float32))
labels.append(int(lf.readline()))
pi+=1
if pi>=10000000:
print 'output {} part {}'.format(fn,part_id)
pi=0
save_room_pkl('data/Semantic3D.Net/pkl/train2/'+fn+'_{}.pkl'.format(part_id),
np.asarray(points,np.float32),np.asarray(labels,np.int32))
points,labels=[],[]
part_id+=1
lf.close()
pf.close()
if pi!=0:
save_room_pkl('data/Semantic3D.Net/pkl/train2/'+fn+'_{}.pkl'.format(part_id),
np.asarray(points,np.float32),np.asarray(labels,np.int32))
print '{} done'.format(fn)
def downsample_and_save():
train_list,test_list=get_block_train_test_split()
idx=0
train_list+=test_list
with open('cached/room_block_ds0.03_stems.txt','w') as f:
for _ in xrange(5):
for fs in train_list:
points,labels=read_room_pkl('data/S3DIS/room_block_10_10/'+fs)
points, labels=downsample(points,labels,0.03)
names=fs.split('_')
names[0]=str(idx)
nfs='_'.join(names)
f.write(nfs+'\n')
ofs='data/S3DIS/room_block_10_10_ds0.03/'+nfs
save_room_pkl(ofs,points,labels)
idx+=1
def prepare_semantic3d_test_partition():
with open('cached/semantic3d_test_stems.txt','r') as f:
fns=f.readlines()
fns=[fn.strip('\n').split(' ')[0] for fn in fns]
for fn in fns:
pf=open('data/Semantic3D.Net/raw/test/' + fn + '.txt')
count=0
while True:
line=pf.readline()
if not line:
break
count+=1
pf.seek(0)
points,labels=[],[]
while True:
line=pf.readline()
if not line:
break
line=line.strip('\n')
sublines=line.split(' ')
x=float(sublines[0])
y=float(sublines[1])
z=float(sublines[2])
intensity=float(sublines[3])
r=float(sublines[4])
g=float(sublines[5])
b=float(sublines[6])
points.append(np.asarray([x,y,z,r,g,b,intensity],dtype=np.float32))
labels.append(0)
pf.close()
save_room_pkl('data/Semantic3D.Net/pkl/test/'+fn+'.pkl',
np.asarray(points,np.float32),np.asarray(labels,np.int32))
print '{} done'.format(fn)