File size: 2,455 Bytes
09f3dee |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 |
from __future__ import print_function, division
import argparse
import os
import copy
import numpy as np
from PIL import Image
import torch
import torch.nn as nn
import torch.nn.functional as F
import re
import glob
import torchvision
import cv2
DEVICE = 'cuda'
maindir = 'path to the selected kitti 2015 dataset'
datasetName = ["1_KITTI"]
# csvprename={'KITTI', 'vkitti','MPI','VIPER','Spring','Monkaa','MHOF','Driving','FT3D' ,'TartanAir'};
datasetN = len(datasetName)
sessionN = 12
movN = 2
frameN = 15
def load_image(imfile):
img = np.array(Image.open(imfile)).astype(np.uint8)
if len(img.shape) == 2:
img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
cv2.imshow('image', img[:, :, [2, 1, 0]] / 255.0)
img = torch.from_numpy(img).permute(2, 0, 1).float()
return img[None].to(DEVICE)
def save_video(flo, img, writer):
# map flow to rgb image
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
flo = cv2.cvtColor(flo, cv2.COLOR_BGR2RGB)
print(flo.shape)
img_flo = np.concatenate([img, flo], axis=0).astype(np.uint8)
writer.write(img_flo)
def demo(args):
for dataset in range(datasetN): # ten dataset in total
for session in range(1, sessionN + 1):
destination_folder = os.path.join(maindir, datasetName[dataset], f'session{session:03d}')
video_file = os.path.join(destination_folder, f'session{session:03d}.mp4')
out = None
for file in glob.glob(os.path.join(destination_folder, 'flow_*.mat')):
os.remove(file)
for mov in range(1, movN + 1):
image_list_ = glob.glob(os.path.join(destination_folder, f'Mov{mov}_F*.jpg'))
if len(image_list_) == 0:
image_list_ = glob.glob(os.path.join(destination_folder, f'Mov{mov}_F*.png'))
image_list_.sort(key=lambda x: int(re.sub('\D', '', x)))
print(image_list_)
# load all images
image_list = [load_image(img) for img in image_list_]
# resize the image to that divisible by 8
image_size_ori = image_list[0].shape[-2:]
image_size = [(image_size_ori[0] // 8 + 1) * 8, (image_size_ori[1] // 8 + 1) * 8]
image_list_resize = [F.interpolate(img, size=image_size, mode='bicubic', align_corners=True) for img in
image_list]
if __name__ == '__main__':
demo()
|