### Python 使用PIL或OpenCV将图片粘贴到另一张完整图片 #### 使用PIL库进行图片粘贴 通过`Pillow`库中的`paste()`方法能够轻松地把一张图片粘贴至另一张图片之上。此过程允许指定位置以及可选的蒙版来控制透明度效果。 ```python from PIL import Image def paste_image_pil(source_img_path, target_img_path, output_path, position=(0, 0)): source = Image.open(source_img_path).convert("RGBA") # 转换为带Alpha通道的模式以便支持半透明 target = Image.open(target_img_path).convert("RGBA") # 创建一个新的alpha掩码，这里简单设置全白即完全不透明 alpha_mask = Image.new('L', source.size, color=255) # 将source image按照给定position粘贴到target image上，并应用alpha mask target.paste(source, box=position, mask=alpha_mask) # 保存最终结果 target.save(output_path) ``` 上述代码展示了如何利用`PIL.Image.paste()`函数并传递适当参数[^4]。 #### 使用OpenCV库进行图片粘贴 对于更复杂的需求比如带有透明背景的情况，则可能需要用到`OpenCV`来进行处理。下面的例子说明了怎样使用`cv2.addWeighted()`或者直接修改ROI区域的方式实现两张图片之间的叠加： ```python import cv2 import numpy as np def overlay_transparent(background_img, img_to_overlay_t, x, y, overlay_size=None): """ @brief Overlays a transparant PNG onto background image at the specified offset. :param background_img: The background image to be overlaid on top of. :param img_to_overlay_t: The transparent image that needs to be placed over `background_img`. :param x,y: Location coordinates where you want your second image's upper-left corner positioned relative to first one. :param overlay_size: Tuple containing desired dimensions for resized version of `img_to_overlay_t`. If None then original size will remain unchanged. Returns modified copy of inputted `background_img`. """ bg_img = background_img.copy() try: # Resize if necessary if overlay_size is not None: img_to_overlay_t = cv2.resize(img_to_overlay_t, dsize=overlay_size) b,g,r,a = cv2.split(img_to_overlay_t) overlay_color = cv2.merge((b,g,r)) h, w, _ = overlay_color.shape roi = bg_img[y:y+h, x:x+w] # Create an inverted mask from alpha channel and apply it to ROI area mask = cv2.threshold(a, 1, 255, cv2.THRESH_BINARY)[1] inv_mask = cv2.bitwise_not(mask) masked_bg = cv2.bitwise_and(roi, roi, mask=inv_mask) final_roi = cv2.add(masked_bg, overlay_color) bg_img[y:y+h, x:x+w] = final_roi return bg_img except Exception as e: print(f"Error occurred during processing images {e}") raise ``` 这段代码实现了更加精细的操作——不仅考虑到了图像大小调整的可能性，还特别针对具有透明背景的对象进行了特殊处理[^1]。