Python+Qt身体特征识别人数统计源码窗体程序(使用步骤)

前言

这篇博客针对《PPython+Qt身体特征识别人数统计》编写代码，功能包括了相片，摄像头身体识别，数量统计。代码整洁，规则，易读。应用推荐首选。

一、所需工具软件          

 1. Python3.6以上           

2. Pycharm代码编辑器          

3. PyQt, Torch库

二、使用步骤

1.引入库

代码如下（示例）：

import cv2
import torch
from numpy import random

from models.experimental import attempt_load
from utils.datasets import LoadStreams, LoadImages
from utils.general import check_img_size, check_requirements, check_imshow, non_max_suppression, apply_classifier, \
   scale_coords, xyxy2xywh, strip_optimizer, set_logging, increment_path
from utils.plots import plot_one_box
from utils.torch_utils import select_device, load_classifier, time_synchronized

2.识别特征图像

代码如下（示例）：

def detect(save_img=False):
   source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
   webcam = source.isnumeric() or source.endswith('.txt') or source.lower().startswith(
       ('rtsp://', 'rtmp://', 'http://'))

# Directories
   save_dir = Path(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok))  # increment run
   (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True)  # make dir

# Initialize
   set_logging()
   device = select_device(opt.device)
   half = device.type != 'cpu'  # half precision only supported on CUDA

# Load model
   model = attempt_load(weights, map_location=device)  # load FP32 model
   stride = int(model.stride.max())  # model stride
   imgsz = check_img_size(imgsz, s=stride)  # check img_size
   if half:
       model.half()  # to FP16

# Second-stage classifier
   classify = False
   if classify:
       modelc = load_classifier(name='resnet101', n=2)  # initialize
       modelc.load_state_dict(torch.load('weights/resnet101.pt', map_location=device)['model']).to(device).eval()

# Set Dataloader
   vid_path, vid_writer = None, None
   if webcam:
       view_img = check_imshow()
       cudnn.benchmark = True  # set True to speed up constant image size inference
       dataset = LoadStreams(source, img_size=imgsz, stride=stride)
   else:
       save_img = True
       dataset = LoadImages(source, img_size=imgsz, stride=stride)

# Get names and colors
   names = model.module.names if hasattr(model, 'module') else model.names
   colors = [[random.randint(0, 255) for _ in range(3)] for _ in names]

# Run inference
   if device.type != 'cpu':
       model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters())))  # run once
   t0 = time.time()
   for path, img, im0s, vid_cap in dataset:
       img = torch.from_numpy(img).to(device)
       img = img.half() if half else img.float()  # uint8 to fp16/32
       img /= 255.0  # 0 - 255 to 0.0 - 1.0
       if img.ndimension() == 3:
           img = img.unsqueeze(0)

# Inference
       t1 = time_synchronized()
       pred = model(img, augment=opt.augment)[0]

# Apply NMS
       pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms)
       t2 = time_synchronized()

# Apply Classifier
       if classify:
           pred = apply_classifier(pred, modelc, img, im0s)

# Process detections
       for i, det in enumerate(pred):  # detections per image
           if webcam:  # batch_size >= 1
               p, s, im0, frame = path[i], '％g: ' ％ i, im0s[i].copy(), dataset.count
           else:
               p, s, im0, frame = path, '', im0s, getattr(dataset, 'frame', 0)

p = Path(p)  # to Path
           save_path = str(save_dir / p.name)  # img.jpg
           txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}')  # img.txt
           s += '％gx％g ' ％ img.shape[2:]  # print string
           gn = torch.tensor(im0.shape)[[1, 0, 1, 0]]  # normalization gain whwh
           if len(det):
               # Rescale boxes from img_size to im0 size
               det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()

# 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 opt.save_conf else (cls, *xywh)  # label format
                       with open(txt_path + '.txt', 'a') as f:
                           f.write(('％g ' * len(line)).rstrip() ％ line + '\n')

if save_img or view_img:  # Add bbox to image
                       label = f'{names[int(cls)]} {conf:.2f}'
                       plot_one_box(xyxy, im0, label=label, color=colors[int(cls)], line_thickness=3)

# Print time (inference + NMS)
           print(f'{s}Done. ({t2 - t1:.3f}s)')

# Save results (image with detections)
           if save_img:
               if dataset.mode == 'image':
                   cv2.imwrite(save_path, im0)
               else:  # 'video'
                   if vid_path != save_path:  # new video
                       vid_path = save_path
                       if isinstance(vid_writer, cv2.VideoWriter):
                           vid_writer.release()  # release previous video writer

fourcc = 'mp4v'  # output video codec
                       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))
                       vid_writer = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*fourcc), fps, (w, h))
                   vid_writer.write(im0)

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 ''
       print(f"Results saved to {save_dir}{s}")

print(f'Done. ({time.time() - t0:.3f}s)')

print(opt)
   check_requirements()

with torch.no_grad():
       if opt.update:  # update all models (to fix SourceChangeWarning)
           for opt.weights in ['yolov5s.pt', 'yolov5m.pt', 'yolov5l.pt', 'yolov5x.pt']:
               detect()
               strip_optimizer(opt.weights)
       else:
           detect()

3.识别参数定义：

代码如下（示例）：

if __name__ == '__main__':
   parser = argparse.ArgumentParser()
   parser.add_argument('--weights', nargs='+', type=str, default='yolov5_best_road_crack_recog.pt', help='model.pt path(s)')
   parser.add_argument('--img-size', type=int, default=640, help='inference size (pixels)')
   parser.add_argument('--conf-thres', type=float, default=0.25, help='object confidence threshold')
   parser.add_argument('--iou-thres', type=float, default=0.45, help='IOU threshold for NMS')
   parser.add_argument('--view-img', action='store_true', help='display results')
   parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
   parser.add_argument('--classes', nargs='+', type=int, default='0', help='filter by class: --class 0, or --class 0 2 3')
   parser.add_argument('--agnostic-nms', action='store_true', help='class-agnostic NMS')
   parser.add_argument('--augment', action='store_true', help='augmented inference')
   parser.add_argument('--update', action='store_true', help='update all models')
   parser.add_argument('--project', default='runs/detect', help='save results to project/name')
   parser.add_argument('--name', default='exp', help='save results to project/name')
   parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
   opt = parser.parse_args()

print(opt)
   check_requirements()

with torch.no_grad():
       if opt.update:  # update all models (to fix SourceChangeWarning)
           for opt.weights in ['yolov5s.pt', 'yolov5m.pt', 'yolov5l.pt', 'yolov5x.pt']:
               detect()
               strip_optimizer(opt.weights)
       else:
           detect()

4.运行结果如下： 

来源：https://blog.csdn.net/alicema1111/article/details/128373886