项目地址：https://github.com/cfzd/Ultra-Fast-Lane-Detection
issue中作者回答了很多训练数据集及检测其他视频的回答

作者知乎文章：https://zhuanlan.zhihu.com/p/157530787

仅修改demo程序以适用于1920*1080分辨率视频。测试使用，如需更好效果应修改源程序并重新训练模型。

1.修改 demo.py

import torch, os
import cv2
from model.model import parsingNet
from utils.common import merge_config
from utils.dist_utils import dist_print
import torch
import scipy.special, tqdm
import numpy as np
import torchvision.transforms as transforms
from data.dataset import LaneTestDataset
from data.constant import culane_row_anchor, tusimple_row_anchorif __name__ == "__main__":torch.backends.cudnn.benchmark = Trueargs, cfg = merge_config()dist_print('start testing...')assert cfg.backbone in ['18','34','50','101','152','50next','101next','50wide','101wide']if cfg.dataset == 'CULane':cls_num_per_lane = 18elif cfg.dataset == 'Tusimple':cls_num_per_lane = 56else:raise NotImplementedErrornet = parsingNet(pretrained = False, backbone=cfg.backbone,cls_dim = (cfg.griding_num+1,cls_num_per_lane,4),use_aux=False).cuda() # we dont need auxiliary segmentation in testingstate_dict = torch.load(cfg.test_model, map_location='cpu')['model']compatible_state_dict = {}for k, v in state_dict.items():if 'module.' in k:compatible_state_dict[k[7:]] = velse:compatible_state_dict[k] = vnet.load_state_dict(compatible_state_dict, strict=False)net.eval()img_transforms = transforms.Compose([#transforms.CenterCrop((590,1640)),transforms.Resize((288, 800)),transforms.ToTensor(),transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),])cap = cv2.VideoCapture("your-test-video.mp4")fourcc = cv2.VideoWriter_fourcc(*'MJPG')rval,frame = cap.read()frame = frame[490:1080,0:1640,:]vout = cv2.VideoWriter('output.avi', fourcc , 30.0, (frame.shape[1],frame.shape[0]))print("w= {},h = {}".format(cap.get(3),cap.get(4)))from PIL import Imageprint('加载CUDA是否成功:',torch.cuda.is_available())while 1:rval,frame = cap.read()if rval == False:breakframe = frame[490:1080,0:1640,:]img = cv2.cvtColor(frame,cv2.COLOR_BGR2RGB)img_ = Image.fromarray(img)imgs = img_transforms(img_)imgs = imgs.unsqueeze(0)imgs = imgs.cuda()with torch.no_grad():out = net(imgs)col_sample = np.linspace(0, 800 - 1, cfg.griding_num)col_sample_w = col_sample[1] - col_sample[0]out_j = out[0].data.cpu().numpy()out_j = out_j[:, ::-1, :]prob = scipy.special.softmax(out_j[:-1, :, :], axis=0)idx = np.arange(cfg.griding_num) + 1idx = idx.reshape(-1, 1, 1)loc = np.sum(prob * idx, axis=0)out_j = np.argmax(out_j, axis=0)loc[out_j == cfg.griding_num] = 0out_j = locfor i in range(out_j.shape[1]):if np.sum(out_j[:, i] != 0) > 2:for k in range(out_j.shape[0]):if out_j[k, i] > 0:ppp = (int(out_j[k, i] * col_sample_w * frame.shape[1] / 800) - 1, int(frame.shape[0] - k * 20) - 1)cv2.circle(frame,ppp,5,(0,255,0),-1)vout.write(frame)vout.release()

2.测试

python demo.py configs/culane.py --test_model path_to_culane_18.pth