yolo v5 、ffmpge、nginx 视频检测
一、ffmpeg 下载
windows: https://www.gyan.dev/ffmpeg/builds/
ubuntu: https://github.com/BtbN/FFmpeg-Builds/releases
ffmpeg-master-latest-linux64-gpl.tar.xz
#新建ffmpeg 文件夹 更改为777 的权限
#解压即可
tar -xf ffmpeg-master-latest-linux64-gpl.tar.xz -C /usr/local/safeware/ffmpeg/ffmpeg --strip-components=1
#检测是否安装好
./ffmpeg -version
二、yolov5 中主要代码
yoloV5的utils/ffmpeg/ffmpeg.py 中创建 ffmpeg.py
复制代码
# subprocess 模块允许我们启动一个新进程，并连接到它们的输入/输出/错误管道，从而获取返回值。
import subprocess
import os
import signal
class FFmpeg():
#fps 帧率 video_size 视频尺寸 push_url 推流到的服务器
def __init__(self, fps, video_size,push_url):
# 创建FFmpeg命令行参数
ffmpeg_cmd = ['F:/chen/ffmpeg-8.1.1-full_build/bin/ffmpeg.exe', #ffmpeg 所在目录
'-itsoffset', '15',
'-r', str(fps[0]), # '24.0' #帧率
'-y', # 覆盖已存在的文件
'-f', 'rawvideo',
'-pixel_format', 'bgr24',
'-thread_queue_size', '1024',
'-video_size', video_size,
'-i', '-', # 从标准输入读取数据
'-c:v', 'libx264',
'-preset', 'ultrafast',
'-tune', 'zerolatency',
'-pix_fmt', 'yuv420p',
'-f', 'flv',
push_url]
print('ffmpeg_cmd:', ffmpeg_cmd)
# 启动 ffmpeg
self.ffmepg_process = subprocess.Popen(ffmpeg_cmd, stdin=subprocess.PIPE)
#开始推流
def streamPush(self, frame):
self.ffmepg_process.stdin.write(frame.tobytes())
#self.ffmepg_process.stdin.write(frame)
#停止
def stopPush(self):
os.kill(self.ffmepg_process.pid, signal.SIGINT)
self.ffmepg_process.wait()改造detect 支持视频检测和推流
创建（模板来自根目录下的 detect.py） myDetect.py
# 自定义调用检测
import base64
import os
import sys
import threading
from pathlib import Path
import cv2
import torch
import torch.backends.cudnn as cudnn
from ultralytics.utils.ops import scale_coords
from ultralytics.utils.plotting import Annotator, save_one_box, colors
from utils.ffmpeg.ffmpeg import FFmpeg
FILE = Path(__file__).resolve()
ROOT = FILE.parents[0] # YOLOv5 root directory
if str(ROOT) not in sys.path:
sys.path.append(str(ROOT)) # add ROOT to PATH
ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
from models.common import DetectMultiBackend
from utils.dataloaders import IMG_FORMATS, VID_FORMATS, LoadImages, LoadScreenshots, LoadStreams
from utils.general import (
LOGGER,
Profile,
check_file,
check_img_size,
check_imshow,
check_requirements,
colorstr,
cv2,
increment_path,
non_max_suppression,
print_args,
scale_boxes,
strip_optimizer,
xyxy2xywh,
)
from utils.torch_utils import select_device, time_sync
class MyDelect():
def __init__(self, weights, source,returnStr):
super().__init__()
self.weights = weights
self.source = source
self.pushStream = None
# 图片地址
self.picPath = ""
self.flag = False
self.returnStr =returnStr
# 真正运行的yolov5代码，加载模型
@torch.no_grad()
def loadModel(self, weights=ROOT / 'runs/train/exp39/weights/best.pt', # 权重模型
source=ROOT / 'data/images/zidane.jpg', # 需要检测的图片地址
imgsz=640, # inference size (pixels) 检测裁剪大小
conf_thres=0.2, # confidence threshold 预测概率大于0.25会被选中
iou_thres=0.45, # NMS IOU threshold 保留的选中框数量
max_det=1000, # maximum detections per image 检测个数
device='0', # cuda device, i.e. 0 or 0,1,2,3 or cpu 检测的运算方式，cpu 还是GPU
view_img=False, # show results 实时输出检测结果
save_txt=False, # save results to *.txt 将检测结果保存
save_conf=False, # save confidences in --save-txt labels 置信度保存，与txt一起使用
save_crop=False, # save cropped prediction boxes 裁剪模型检测到的物体
nosave=False, # do not save images/videos 是否保存预测结果
classes=None, # filter by class: --class 0, or --class 0 2 3 类别数量
agnostic_nms=False, # class-agnostic NMS 同一个物体检测为多个分类是否全显示选中框
augment=False, # augmented inference 提高预测率
visualize=False, # visualize features 保存特征文件，灰阶图
update=False, # update all models
project=ROOT / 'runs/detect', # save results to project/name 预测结果保存路径
name='exp', # save results to project/name 保存的路径名字
exist_ok=False, # existing project/name ok, do not increment 覆盖保存，检测结果都在相同文件夹下
line_thickness=1, # bounding box thickness (pixels) #标注框细腻的
hide_labels=False, # hide labels 是否隐藏标签
hide_conf=False, # hide confidences 隐藏置信度
half=False, # use FP16 half-precision inference 低精度预测，用于设备性能较弱的设备
dnn=False # use OpenCV DNN for ONNX inference 使用openCV 的DNN进行ONNX推理
):
source = self.source
weights = self.weights
source = str(source)
# 将source 转字符
save_img = not nosave and not source.endswith('.txt') # save inference images 保存检测结果
is_file = Path(source).suffix[1:] in (IMG_FORMATS + VID_FORMATS) # 判断sourc的格式，文件格式
is_url = source.lower().startswith(('rtsp://', 'rtmp://', 'http://', 'https://')) # 网络资源
webcam = source.isnumeric() or source.endswith('.txt') or (is_url and not is_file) # 当source 是视频将逐帧检测
# 判断是否是视频
# 将文件下载保存
if is_url and is_file:
source = check_file(source) # download
# Directories 保存文件夹，结果
save_dir = increment_path(Path(project) / name, exist_ok=exist_ok) # increment run
(save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir
# Load model 模型加载
device = select_device(device)
# 视频及摄像头调用模型
model = DetectMultiBackend(self.weights, device=device, dnn=dnn)
stride, names, pt, jit, onnx = model.stride, model.names, model.pt, model.jit, model.onnx
imgsz = [640, 640] # 设置为数组
imgsz = check_img_size(imgsz, s=stride) # check image size
# Half
half &= pt and device.type != 'cpu' # half precision only supported by PyTorch on CUDA
if pt:
model.model.half() if half else model.model.float()
# Dataloader 加载资源
if webcam:
view_img = check_imshow()
cudnn.benchmark = True # set True to speed up constant image size inference
# 调用摄像头
dataset = LoadStreams(source, img_size=imgsz, stride=stride, auto=pt and not jit)
bs = len(dataset) # batch_size
else:
dataset = LoadImages(source, img_size=imgsz, stride=stride, auto=pt and not jit)
bs = 1 # batch_size
vid_path, vid_writer = [None] * bs, [None] * bs # vid_path 和 vid_writer 分别是视频路径和视频编写器
# Run inference
if pt and device.type != 'cpu':
model(torch.zeros(1, 3, *imgsz).to(device).type_as(next(model.model.parameters()))) # warmup
dt, seen = [0.0, 0.0, 0.0], 0
# 创建Ffmepg对象
if (self.pushStream == None):
videoSize = str(dataset.my_w) + "x"+str(dataset.my_h)
self.pushStream = FFmpeg( dataset.fps,videoSize,"rtmp://192.168.2.136/live/test")
for path, im, im0s, vid_cap, s in dataset:
if (self.flag):
if (self.source == '0'):
# 关闭摄像头
dataset.closeCamer()
return
print("stop")
return
t1 = time_sync()
im = torch.from_numpy(im).to(device)
im = im.half() if half else im.float() # uint8 to fp16/32
im /= 255 # 0 - 255 to 0.0 - 1.0
if len(im.shape) == 3:
im = im[None] # expand for batch dim
t2 = time_sync()
dt[0] += t2 - t1
# Inference
visualize = increment_path(save_dir / Path(path).stem, mkdir=True) if visualize else False
pred = model(im, augment=augment, visualize=visualize)
t3 = time_sync()
dt[1] += t3 - t2
# NMS
pred = non_max_suppression(pred, conf_thres, iou_thres, classes, agnostic_nms, max_det=max_det)
dt[2] += time_sync() - t3
# Second-stage classifier (optional)
# pred = utils.general.apply_classifier(pred, classifier_model, im, im0s)
# Process predictions
for i, det in enumerate(pred): # per image
seen += 1
if webcam: # batch_size >= 1
p, im0, frame = path[i], im0s[i].copy(), dataset.count
s += f'{i}: '
else:
p, im0, frame = path, im0s.copy(), getattr(dataset, 'frame', 0)
p = Path(p) # to Path
save_path = str(save_dir / p.name) # im.jpg
self.picPath = save_path
txt_path = str(save_dir / 'labels' / p.stem) + (
'' if dataset.mode == 'image' else f'_{frame}') # im.txt
s += '%gx%g ' % im.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] # normalization gain whwh
imc = im0.copy() if save_crop else im0 # for save_crop
annotator = Annotator(im0, line_width=line_thickness, example=str(names))
if len(det):
count = 0#分类计数
# Rescale boxes from img_size to im0 size
det[:, :4] = scale_coords(im.shape[2:], det[:, :4], im0.shape).round()
# Print results
for c in det[:, -1].unique():
n = (det[:, -1] == c).sum() # detections per class
s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
# Write results
for *xyxy, conf, cls in reversed(det):
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
line = (cls, *xywh, conf) if save_conf else (cls, *xywh) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * len(line)).rstrip() % line + '\n')
test = ('%g ' * len(line)).rstrip() % line + '\n'
# print(test.split()[len(test.split())-1])
if (cls == 0):
count = count + 1
if save_img or save_crop or view_img: # Add bbox to image
c = int(cls) # integer class
label = None if hide_labels else (names[c] if hide_conf else f'{names[c]} {conf:.2f}')
annotator.box_label(xyxy, label, color=colors(c, True))
if save_crop:
save_one_box(xyxy, imc, file=save_dir / 'crops' / names[c] / f'{p.stem}.jpg', BGR=True)
# Print time (inference-only)
# LOGGER.info(f'{s}Done. ({t3 - t2:.3f}s)')
# Stream results
im0 = annotator.result()
if source.endswith("mp4"):
view_img = True
# 打开视频播放窗口，推流到ffmpeg服务器
if view_img:
# text = 'person:%d ' % (count)
# cv2.putText(im0, text, (180, 50), cv2.FONT_HERSHEY_SIMPLEX, 2, (0, 0, 255), 5)
# 推流
self.pushStream.streamPush(im0)
# 打开播放窗口
cv2.imshow(str(p), im0)
cv2.waitKey(1) # 1 millisecond
# Save results (image with detections)
if save_img:
if dataset.mode == 'image':
# 图片保存
cv2.imwrite(save_path, im0)
else: # 'video' or 'stream'
# 视频
if vid_path[i] != save_path: # new video
vid_path[i] = save_path
if isinstance(vid_writer[i], cv2.VideoWriter):
vid_writer[i].release() # release previous video writer
if vid_cap: # video
fps = vid_cap.get(cv2.CAP_PROP_FPS)
w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
else: # stream
fps, w, h = 30, im0.shape[1], im0.shape[0]
save_path += '.mp4'
vid_writer[i] = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (w, h))
vid_writer[i].write(im0)
# Print results
t = tuple(x / seen * 1E3 for x in dt) # speeds per image
LOGGER.info(f'Speed: %.1fms pre-process, %.1fms inference, %.1fms NMS per image at shape {(1, 3, *imgsz)}' % t)
if save_txt or save_img:
s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}{s}")
if update:
strip_optimizer(weights) # update model (to fix SourceChangeWarning)
return self.picPath
# 视频
def useYolo(self):
if (self.flag == True):
self.flag = False
self.thread = threading.Thread(target=self.loadModel)
self.thread.start()
def useYoloImg(self):
# 将图片转换成base64
result = self.loadModel()
with open(result, 'rb') as img:
# 使用base64进行编码
b64encode = base64.b64encode(img.read())
s = b64encode.decode()
b64_encode = 'data:image/jpeg;base64,%s' % s
# 返回base64编码字符串
return b64_encode
def stopYolo(self):
self.flag = True
self.thread.join()
self.pushStream.stopPush()
print("Thread stopped")测试代码 test.py
def test():
myDetect = MyDelect("F:\\chen\\yolov5\\runs\\train\\exp3\\weights\\best.pt",
"http://60.10.139.113:8801/hls/13/index.m3u8","")
myDetect.useYolo()
if __name__ == "__main__":
test()使用VLC进行串流播放，nginx串流地址: rtmp://192.168.2.136/live/test，结果如下
注：yolov5 是一帧一帧的检测修改的，所以需要将流修改后再推到推流服务器上，这里使用的是nginx搭建的推流服务器，搭建参考 ：http://8.137.94.77:7090/archives/nginx-tui-liu-la-liu-fu-wu-qi
目前推流的视频在播放时存在卡顿的现象，需要进一步优化
可能是目前电脑配置不够（逐帧处理返回速度导致的）或者是ffmpeg推流配置没正确处理