Upload 2 files

image2vidimg.py

@@ -0,0 +1,141 @@
+import torch
+from PIL import Image
+import matplotlib.pyplot as plt
+from extract.getim import load_image
+from torchvision import transforms
+import cv2
+import matplotlib.pyplot as plt
+
+
+
+transform = transforms.Compose([
+    transforms.ToTensor(),  # 将numpy数组或PIL.Image读的图片转换成(C,H, W)的Tensor格式且/255归一化到[0,1.0]之间
+])  # 来自ImageNet的mean和variance
+
+
+# fcontent = load_image("./ori/0.jpg",transform=None,shape=[512, 256])
+
+
+
+def show_cut(path, left, upper, right, lower):
+    """
+        原图与所截区域相比较
+    :param path: 图片路径
+    :param left: 区块左上角位置的像素点离图片左边界的距离
+    :param upper：区块左上角位置的像素点离图片上边界的距离
+    :param right：区块右下角位置的像素点离图片左边界的距离
+    :param lower：区块右下角位置的像素点离图片上边界的距离
+     故需满足：lower > upper、right > left
+    """
+
+    img = path
+
+    # print("This image's size: {}".format(img.size))  # (W, H)
+    # img.save("kkk.jpg")
+    # plt.figure("Image Contrast")
+    #
+    # plt.subplot(1, 2, 1)
+    # plt.title('origin')
+    #
+    # plt.imshow(img)
+    # plt.axis('off')
+    #
+    # box = (left, upper, right, lower)
+    # roi = img.crop(box)
+    #
+    # plt.subplot(1, 2, 2)
+    # plt.title('roi')
+    # plt.imshow(roi)
+    # plt.axis('off')
+    # plt.show()
+
+
+def image_cut_save(path, left, upper, right, lower):
+    """
+        所截区域图片保存
+    :param path: 图片路径
+    :param left: 区块左上角位置的像素点离图片左边界的距离
+    :param upper：区块左上角位置的像素点离图片上边界的距离
+    :param right：区块右下角位置的像素点离图片左边界的距离
+    :param lower：区块右下角位置的像素点离图片上边界的距离
+     故需满足：lower > upper、right > left
+    :param save_path: 所截图片保存位置
+    """
+    img = path  # 打开图像
+    box = (left, upper, right, lower)
+    roi = img.crop(box)
+    # roi.save(save_path)
+    return transform(roi)
+
+
+    # 保存截取的图片
+
+
+
+# def getcontent(fcontent,gap):
+#     Intgap=gap/9
+#     a=torch.Tensor()
+#     for i in range(10):
+#         pic_path = fcontent
+#         # pic_save_dir_path = './out2/0-'+str(i)+".jpg"
+#         left, upper, right, lower = Intgap*i, 0, Intgap*i+gap, gap
+#         a=torch.cat([a,image_cut_save(pic_path, left, upper, right, lower).unsqueeze(1)],dim=1)
+#     return a
+# def cobtwoten(image_path):
+#     fcontent = load_image(image_path, transform=None, shape=[512, 256])
+#     Intgap = 256
+#     a = torch.Tensor()
+#     for i in range(2):
+#         pic_path = fcontent
+#         #pic_save_dir_path = './out2/0-' + str(i) + ".jpg"
+#         left, upper, right, lower = Intgap * i, 0, Intgap * i + Intgap, Intgap
+#         a = torch.cat([a, image_cut_save(pic_path, left, upper, right, lower).unsqueeze(1)], dim=1)
+#     return a.unsqueeze(0)
+
+def cobtwoten(image_path):
+    fcontent = load_image(image_path, transform=None, shape=[256, 128])
+    Intgap = 128/9
+    a = torch.Tensor()
+    for i in range(10):
+        pic_path = fcontent
+        #pic_save_dir_path = './out2/0-' + str(i) + ".jpg"
+        left, upper, right, lower = Intgap * i, 0, Intgap * i + 128, 128
+        a = torch.cat([a, image_cut_save(pic_path, left, upper, right, lower).unsqueeze(1)], dim=1)
+    return a.unsqueeze(0)
+
+def cobtwoten256(image_path):
+    fcontent = load_image(image_path, transform=None, shape=[512,256])
+    Intgap = 256/9
+    a = torch.Tensor()
+    for i in range(10):
+        pic_path = fcontent
+        #pic_save_dir_path = './out2/0-' + str(i) + ".jpg"
+        left, upper, right, lower = Intgap * i, 0, Intgap * i + 256, 256
+        a = torch.cat([a, image_cut_save(pic_path, left, upper, right, lower).unsqueeze(1)], dim=1)
+    return a.unsqueeze(0)
+
+#
+# fcontent = load_image("./extract/image/0.jpg",transform=None,shape=[256,128])
+# Intgap = 128
+# a = torch.Tensor()
+# for i in range(2):
+#     pic_path = fcontent
+#     pic_save_dir_path = './out2/0-'+str(i)+".jpg"
+#     left, upper, right, lower = Intgap * i, 0, Intgap * i + 128, 128
+#     a = torch.cat([a, image_cut_save(pic_path, left, upper, right, lower,pic_save_dir_path).unsqueeze(1)], dim=1)
+# print(a.shape)
+import numpy as np
+def imgsave(image, path):
+    image = image.squeeze(0)
+    image = image.permute(1, 2, 0)
+    image_np = image.cpu().numpy()*255
+    image_np = image_np.astype(np.uint8)
+    Image.fromarray(image_np).save(path)  # 直接保存PIL图像对象
+
+# lik=["0"]
+# for name in lik:
+#     videos=cobtwoten("./extract/image/0.jpg").permute(0, 2, 1, 3, 4)
+#     print(videos.shape)
+#     for i in range(10):
+#         frame = videos[:, i, :, :]
+#         imgsave(frame, "./out2/"+str(i)+".jpg")

tensor2video.py

@@ -0,0 +1,23 @@
+import os
+import imageio
+from PIL import Image
+
+# 设置生成的视频文件名和路径
+filename = 'output.mp4'
+filepath = os.path.join(os.getcwd(), filename)
+print(filepath)
+
+# 读取所有 PNG 图片
+images = []
+for file_name in sorted(os.listdir("./out2/")):
+    if file_name.endswith('.jpg'):
+        images.append(Image.open("./out2/"+file_name))
+
+
+import numpy as np
+# 将图片转换为视频
+fps = 25  # 每秒钟30帧
+with imageio.get_writer(filepath, fps=fps) as video:
+    for image in images:
+        frame = np.array(image.convert('RGB'))
+        video.append_data(frame)