imagenet_class_index.cs文件下载

https://download.csdn.net/download/qq_42864343/88492936

代码

import osimport numpy as np
import pandas as pdimport cv2 # opencv-python
from tqdm import tqdm # 进度条
from PIL import Image # pillow
import matplotlib.pyplot as plt
%matplotlib inlineimport torch
import torch.nn.functional as F
from torchvision import models
import time

# 有 GPU 就用 GPU，没有就用 CPU
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
print('device:', device)

# 载入预训练图像分类模型
model = models.resnet18(pretrained=True)
model = model.eval()
model = model.to(device)

# 将idx与类名相对应
df = pd.read_csv('data/imagenet_class_index.csv')
idx_to_labels = {}
for idx, row in df.iterrows():# 英文类名idx_to_labels[row['ID']] = row['class']

# 图像预处理
from torchvision import transforms# 测试集图像预处理-RCTN：缩放裁剪、转 Tensor、归一化
test_transform = transforms.Compose([transforms.Resize(256),transforms.CenterCrop(224),transforms.ToTensor(),transforms.Normalize(mean=[0.485, 0.456, 0.406],std=[0.229, 0.224, 0.225])])

# 处理一帧的函数，供后面调用
def process_frame(img):'''输入摄像头拍摄画面bgr-array，输出图像分类预测结果bgr-array'''# 记录该帧开始处理的时间start_time = time.time()img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) # BGR转RGBimg_pil = Image.fromarray(img_rgb) # array 转 PILinput_img = test_transform(img_pil).unsqueeze(0).to(device) # 预处理pred_logits = model(input_img) # 执行前向预测，得到所有类别的 logit 预测分数pred_softmax = F.softmax(pred_logits, dim=1) # 对 logit 分数做 softmax 运算top_n = torch.topk(pred_softmax, 5) # 取置信度最大的 n 个结果pred_ids = top_n[1].cpu().detach().numpy().squeeze() # 解析预测类别confs = top_n[0].cpu().detach().numpy().squeeze() # 解析置信度# 在图像上写字for i in range(len(confs)):pred_class = idx_to_labels[pred_ids[i]]text = '{:<15} {:>.3f}'.format(pred_class, confs[i])# 图片，添加的文字，左上角坐标，字体，字体大小，颜色，线宽，线型img = cv2.putText(img, text, (50, 160 + 80 * i), cv2.FONT_HERSHEY_SIMPLEX, 2, (0, 0, 255), 4, cv2.LINE_AA)# 记录该帧处理完毕的时间end_time = time.time()# 计算每秒处理图像帧数FPSFPS = 1/(end_time - start_time)# 图片，添加的文字，左上角坐标，字体，字体大小，颜色，线宽，线型img = cv2.putText(img, 'FPS  '+str(int(FPS)), (50, 80), cv2.FONT_HERSHEY_SIMPLEX, 2, (255, 0, 255), 4, cv2.LINE_AA)return img

# 调用摄像头处理摄像头中的画面
def view_video(video_path):# 设置显示窗口的大小width,height = 800,600video = cv2.VideoCapture(video_path)'''把摄像头设置为1980 x 1080'''video.set(cv2.CAP_PROP_FRAME_WIDTH,1920)video.set(cv2.CAP_PROP_FRAME_HEIGHT,1080)video.set(cv2.CAP_PROP_FOURCC,cv2.VideoWriter.fourcc('M','J','P','G'))if video.isOpened():'''video.read() 一帧一帧地读取open 得到的是一个布尔值，就是 True 或者 Falseframe 得到当前这一帧的图像'''open, frame = video.read()else:open = Falsewhile open:ret, frame = video.read()# 如果读到的帧数不为空，那么就继续读取，如果为空，就退出if frame is None:breakif ret == True:# !!!处理帧函数frame = process_frame(frame)cv2.namedWindow('video',cv2.WINDOW_NORMAL)cv2.imshow("video", frame)# 50毫秒内判断是否受到esc按键的信息if cv2.waitKey(50) & 0xFF == 27:breakvideo.release()cv2.destroyAllWindows()

# linux usb摄像头的Id一般为1
view_video(1)