### 拉普拉斯金字塔在Python中实现可见光与红外图像融合 拉普拉斯金字塔是一种多分辨率分解方法，常用于图像融合。通过将可见光和红外图像分别分解为多个层次的细节信息，并对这些细节信息进行融合，最终可以得到融合后的图像[^2]。 以下是基于拉普拉斯金字塔的图像融合实现代码示例： ```python import cv2 import numpy as np def laplacian_pyramid(image, levels): """生成拉普拉斯金字塔""" gaussian_pyramid = [image] for _ in range(levels): image = cv2.pyrDown(image) gaussian_pyramid.append(image) laplacian_pyramid = [] for i in range(len(gaussian_pyramid) - 1, 0, -1): size = (gaussian_pyramid[i - 1].shape[1], gaussian_pyramid[i - 1].shape[0]) expanded = cv2.pyrUp(gaussian_pyramid[i], dstsize=size) laplacian = cv2.subtract(gaussian_pyramid[i - 1], expanded) laplacian_pyramid.append(laplacian) laplacian_pyramid.append(gaussian_pyramid[-1]) # 添加顶层 return laplacian_pyramid def fuse_laplacian_pyramids(pyramid1, pyramid2): """融合两个拉普拉斯金字塔""" fused_pyramid = [] for level1, level2 in zip(pyramid1, pyramid2): fused_level = (level1 + level2) / 2 fused_pyramid.append(fused_level) return fused_pyramid def reconstruct_image(laplacian_pyramid): """从拉普拉斯金字塔重构图像""" fused_image = laplacian_pyramid[-1] for i in range(len(laplacian_pyramid) - 2, -1, -1): size = (laplacian_pyramid[i].shape[1], laplacian_pyramid[i].shape[0]) fused_image = cv2.pyrUp(fused_image, dstsize=size) fused_image = cv2.add(fused_image, laplacian_pyramid[i]) return fused_image # 读取可见光和红外图像 visible_image = cv2.imread("visible.jpg", cv2.IMREAD_GRAYSCALE) infrared_image = cv2.imread("infrared.jpg", cv2.IMREAD_GRAYSCALE) # 确保图像尺寸一致 if visible_image.shape != infrared_image.shape: h, w = min(visible_image.shape[0], infrared_image.shape[0]), min(visible_image.shape[1], infrared_image.shape[1]) visible_image = cv2.resize(visible_image, (w, h)) infrared_image = cv2.resize(infrared_image, (w, h)) # 设置金字塔层数 levels = 4 # 构建拉普拉斯金字塔 visible_pyramid = laplacian_pyramid(visible_image, levels) infrared_pyramid = laplacian_pyramid(infrared_image, levels) # 融合拉普拉斯金字塔 fused_pyramid = fuse_laplacian_pyramids(visible_pyramid, infrared_pyramid) # 重构融合图像 fused_image = reconstruct_image(fused_pyramid) # 保存融合结果 cv2.imwrite("fused_image.jpg", fused_image) ``` #### 代码说明 - **`laplacian_pyramid`**：构建拉普拉斯金字塔，输入为图像和金字塔层数。 - **`fuse_laplacian_pyramids`**：将两个拉普拉斯金字塔逐层加权平均融合。 - **`reconstruct_image`**：从拉普拉斯金字塔重构出融合后的图像。 - 图像融合过程中使用了加权平均法[^2]。