import cv2
import torch
import numpy as np
from utils.general import non_max_suppression, scale_coords, xyxy2xywh
from utils.augmentations import letterbox
from models.experimental import attempt_load
import os

weights = r'D:\yolov5-master\runs\train\exp17\weights\best.pt' #已训练好的模型路径

device = 'cuda' if torch.cuda.is_available() else 'cpu'
half = device != 'cpu'
imgsz = 640
conf_thres = 0.2
iou_thres = 0.2
model = attempt_load(weights, map_location=device)  # load FP32 model
stride = int(model.stride.max())  #model stride
names = model.module.names if hasattr(model, 'module') else model.names  #class name
if half:
    model.half()  # to FP16W
if device != 'cpu':
    model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters()))) #run once 初始化





def run():
        entities=os.scandir("E:\Project\Script\pig\images")
        for it in entities:
            img0=cv2.imread(it.path)
            img0 = cv2.cvtColor(img0, cv2.COLOR_BGRA2BGR)
            img = letterbox(img0, imgsz, stride=stride)[0] #调整图片大小，填充
            img = img.transpose((2, 0, 1))[::-1] #HWC to CHW, BGR to RBG
            _,w,h=img.shape
            img = np.ascontiguousarray(img)
            img = torch.from_numpy(img).to(device)
            img = img.half() if half else img.float()
            img /= 255.0
            if len(img.shape) == 3:
                img = img[None]  # expand for batch dim


            pred = model(img, augment=False, visualize=False)[0]
            pred = non_max_suppression(pred, conf_thres, iou_thres, agnostic=False)

            aims = []
            for i, det in enumerate(pred):
                gn = torch.tensor(img0.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], img0.shape).round()

                    for *xyxy, conf, cls in reversed(det):
                        # bbox:(tag, x_center, y_center, x_width, y_width)
                        xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist()  # normalized xywh
                        line = (cls, *xywh)  # label format
                        aim = ('%g ' * len(line)).rstrip() % line
                        aim = aim.split(' ')
                        # aim=list(map(float,aim))
                        aims.append(aim)
            f=open("E:\Project\Script\pig\labels\\"+it.name.split('.')[0]+'.txt','w')
            for i in aims:
                f.write(' '.join(i)+'\n')
            f.close()
            print(it.name)

run()