[已解决]将其中的canny检测算子，更改为sobel检测算子。请在原代码更改谢谢

Justheartyoung

import cv2
import numpy as np
# 1. 分割遥感影像为多块图幅
def split_image(image, block_size):
    height, width = image.shape[:2]
    block_images = []
    for i in range(0, height, block_size):
        for j in range(0, width, block_size):
            block = image[i:i+block_size, j:j+block_size]
            block_images.append(block)
    return block_images

# 2. 分别对每块图幅转为灰度图像
def convert_to_gray(image):
    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    return gray_image

# 3. 对各个图像进行边缘检测
def edge_detection(image):
    edges = cv2.Canny(image, 0.5, 2)
    return edges


# 4. 消除噪声干扰
def remove_noise(image):
    denoised_image = cv2.medianBlur(image, 5)
    return denoised_image

# 5. 将各个图幅合并为一个图幅

def merge_images(images, block_size):
    num_blocks = len(images)
    rows = int(np.ceil(np.sqrt(num_blocks)))  # 向上取整
    cols = rows
    merged_image = np.zeros((rows * block_size, cols * block_size), dtype=np.uint8)
    for i, image in enumerate(images):
        x = (i % cols) * block_size
        y = (i // cols) * block_size
        merged_image[y:y+image.shape[0], x:x+image.shape[1]] = image
    return merged_image

# 6. 导出结果
def export_result(image, file_path):
    cv2.imwrite(file_path, image)

# 示例代码
image_path = r"C:\Users\WINDOWS\Desktop\taiyuan\python\dune\dune.tif"
block_size = 800

# 读取遥感影像
image = cv2.imread(image_path)

# 1. 分割遥感影像为多块图幅
block_images = split_image(image, block_size)

# 2. 分别对每块图幅转为灰度图像
gray_images = [convert_to_gray(block) for block in block_images]

# 3. 对各个图像进行边缘检测
edges_images = [edge_detection(gray) for gray in gray_images]

# 4. 消除噪声干扰
denoised_images = [remove_noise(edges) for edges in edges_images]

# 5. 将各个图幅合并为一个图幅
merged_image = merge_images(denoised_images, block_size)

# 6. 导出结果
export_result(merged_image, "output.jpg")

最佳答案

月排行榜 / 总排行榜

学习编程中的Ben

2023-8-7 16:57:57

import cv2import numpy as np# 1. 分割遥感影像为多块图幅def split_image(image, block_size):    height, width = image.shape[:2]    block_images = []    for i in range(0, height, block_size):        for j in range(0, width, block_size):            block = image[i:i+block_size, j:j+block_size]            block_images.append(block)    return block_images# 2. 分别对每块图幅转为灰度图像def convert_to_gray(image):    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)    return gray_image# 3. 对各个图像进行边缘检测（使用Sobel算子）def edge_detection(image):    sobel_x = cv2.Sobel(image, cv2.CV_64F, 1, 0, ksize=3)    sobel_y = cv2.Sobel(image, cv2.CV_64F, 0, 1, ksize=3)    edges = cv2.magnitude(sobel_x, sobel_y)    edges = cv2.convertScaleAbs(edges)    return edges# 4. 消除噪声干扰def remove_noise(image):    denoised_image = cv2.medianBlur(image, 5)    return denoised_image# 5. 将各个图幅合并为一个图幅def merge_images(images, block_size):    num_blocks = len(images)    rows = int(np.ceil(np.sqrt(num_blocks)))  # 向上取整    cols = rows    merged_image = np.zeros((rows * block_size, cols * block_size), dtype=np.uint8)    for i, image in enumerate(images):        x = (i % cols) * block_size        y = (i // cols) * block_size        merged_image[y:y+image.shape[0], x:x+image.shape[1]] = image    return merged_image# 6. 导出结果def export_result(image, file_path):    cv2.imwrite(file_path, image)# 示例代码image_path = r"C:\Users\WINDOWS\Desktop\taiyuan\python\dune\dune.tif"block_size = 800# 读取遥感影像image = cv2.imread(image_path)# 1. 分割遥感影像为多块图幅block_images = split_image(image, block_size)# 2. 分别对每块图幅转为灰度图像gray_images = [convert_to_gray(block) for block in block_images]# 3. 对各个图像进行边缘检测edges_images = [edge_detection(gray) for gray in gray_images]# 4. 消除噪声干扰denoised_images = [remove_noise(edges) for edges in edges_images]# 5. 将各个图幅合并为一个图幅merged_image = merge_images(denoised_images, block_size)# 6. 导出结果export_result(merged_image, "output.jpg")

注意，在第3步中，我将边缘检测算子从Canny改为了Sobel。使用Sobel算子可以获得基于梯度的边缘信息。

如果回答对你有帮助，请给我一个最佳答案！

跳转到最佳答案楼层

学习编程中的Ben

import cv2import numpy as np# 1. 分割遥感影像为多块图幅def split_image(image, block_size):    height, width = image.shape[:2]    block_images = []    for i in range(0, height, block_size):        for j in range(0, width, block_size):            block = image[i:i+block_size, j:j+block_size]            block_images.append(block)    return block_images# 2. 分别对每块图幅转为灰度图像def convert_to_gray(image):    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)    return gray_image# 3. 对各个图像进行边缘检测（使用Sobel算子）def edge_detection(image):    sobel_x = cv2.Sobel(image, cv2.CV_64F, 1, 0, ksize=3)    sobel_y = cv2.Sobel(image, cv2.CV_64F, 0, 1, ksize=3)    edges = cv2.magnitude(sobel_x, sobel_y)    edges = cv2.convertScaleAbs(edges)    return edges# 4. 消除噪声干扰def remove_noise(image):    denoised_image = cv2.medianBlur(image, 5)    return denoised_image# 5. 将各个图幅合并为一个图幅def merge_images(images, block_size):    num_blocks = len(images)    rows = int(np.ceil(np.sqrt(num_blocks)))  # 向上取整    cols = rows    merged_image = np.zeros((rows * block_size, cols * block_size), dtype=np.uint8)    for i, image in enumerate(images):        x = (i % cols) * block_size        y = (i // cols) * block_size        merged_image[y:y+image.shape[0], x:x+image.shape[1]] = image    return merged_image# 6. 导出结果def export_result(image, file_path):    cv2.imwrite(file_path, image)# 示例代码image_path = r"C:\Users\WINDOWS\Desktop\taiyuan\python\dune\dune.tif"block_size = 800# 读取遥感影像image = cv2.imread(image_path)# 1. 分割遥感影像为多块图幅block_images = split_image(image, block_size)# 2. 分别对每块图幅转为灰度图像gray_images = [convert_to_gray(block) for block in block_images]# 3. 对各个图像进行边缘检测edges_images = [edge_detection(gray) for gray in gray_images]# 4. 消除噪声干扰denoised_images = [remove_noise(edges) for edges in edges_images]# 5. 将各个图幅合并为一个图幅merged_image = merge_images(denoised_images, block_size)# 6. 导出结果export_result(merged_image, "output.jpg")

注意，在第3步中，我将边缘检测算子从Canny改为了Sobel。使用Sobel算子可以获得基于梯度的边缘信息。

如果回答对你有帮助，请给我一个最佳答案！

陶远航

在原代码中，将Canny边缘检测算子更改为Sobel边缘检测算子，请看下面的修改后的代码：

import cv2
import numpy as np

# 1. 分割遥感影像为多块图幅
def split_image(image, block_size):
    height, width = image.shape[:2]
    block_images = []
    for i in range(0, height, block_size):
        for j in range(0, width, block_size):
            block = image[i:i+block_size, j:j+block_size]
            block_images.append(block)
    return block_images

# 2. 分别对每块图幅转为灰度图像
def convert_to_gray(image):
    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    return gray_image

# 3. 对各个图像进行边缘检测
def edge_detection(image):
    sobel_x = cv2.Sobel(image, cv2.CV_64F, 1, 0, ksize=3)
    sobel_y = cv2.Sobel(image, cv2.CV_64F, 0, 1, ksize=3)
    edges = np.sqrt(np.square(sobel_x) + np.square(sobel_y))
    edges = edges.astype(np.uint8)
    return edges

# 4. 消除噪声干扰
def remove_noise(image):
    denoised_image = cv2.medianBlur(image, 5)
    return denoised_image

# 5. 将各个图幅合并为一个图幅
def merge_images(images, block_size):
    num_blocks = len(images)
    rows = int(np.ceil(np.sqrt(num_blocks)))  # 向上取整
    cols = rows
    merged_image = np.zeros((rows * block_size, cols * block_size), dtype=np.uint8)
    for i, image in enumerate(images):
        x = (i % cols) * block_size
        y = (i // cols) * block_size
        merged_image[y:y+image.shape[0], x:x+image.shape[1]] = image
    return merged_image

# 6. 导出结果
def export_result(image, file_path):
    cv2.imwrite(file_path, image)

# 示例代码
image_path = r"C:\Users\WINDOWS\Desktop\taiyuan\python\dune\dune.tif"
block_size = 800

# 读取遥感影像
image = cv2.imread(image_path)

# 1. 分割遥感影像为多块图幅
block_images = split_image(image, block_size)

# 2. 分别对每块图幅转为灰度图像
gray_images = [convert_to_gray(block) for block in block_images]

# 3. 对各个图像进行边缘检测
edges_images = [edge_detection(gray) for gray in gray_images]

# 4. 消除噪声干扰
denoised_images = [remove_noise(edges) for edges in edges_images]

# 5. 将各个图幅合并为一个图幅
merged_image = merge_images(denoised_images, block_size)

# 6. 导出结果
export_result(merged_image, "output.jpg")

这段修改后的代码将使用Sobel边缘检测算子进行图像边缘检测。希望对你有所帮助！如果还有其他问题，请随时提问。
如果问题已经解决，请设置最佳答案

isdkz

import cv2
import numpy as np

# 1. 分割遥感影像为多块图幅
def split_image(image, block_size):
    height, width = image.shape[:2]
    block_images = []
    for i in range(0, height, block_size):
        for j in range(0, width, block_size):
            block = image[i:i+block_size, j:j+block_size]
            block_images.append(block)
    return block_images

# 2. 分别对每块图幅转为灰度图像
def convert_to_gray(image):
    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    return gray_image

# 3. 对各个图像进行边缘检测
def edge_detection(image):
    edges = cv2.Sobel(image, cv2.CV_64F, 1, 1, ksize=3)
    edges = cv2.convertScaleAbs(edges)
    return edges

# 4. 消除噪声干扰
def remove_noise(image):
    denoised_image = cv2.medianBlur(image, 5)
    return denoised_image

# 5. 将各个图幅合并为一个图幅
def merge_images(images, block_size):
    num_blocks = len(images)
    rows = int(np.ceil(np.sqrt(num_blocks)))  # 向上取整
    cols = rows
    merged_image = np.zeros((rows * block_size, cols * block_size), dtype=np.uint8)
    for i, image in enumerate(images):
        x = (i % cols) * block_size
        y = (i // cols) * block_size
        merged_image[y:y+image.shape[0], x:x+image.shape[1]] = image
    return merged_image

# 6. 导出结果
def export_result(image, file_path):
    cv2.imwrite(file_path, image)

# 示例代码
image_path = r"C:\Users\WINDOWS\Desktop\taiyuan\python\dune\dune.tif"
block_size = 800

# 读取遥感影像
image = cv2.imread(image_path)

# 1. 分割遥感影像为多块图幅
block_images = split_image(image, block_size)

# 2. 分别对每块图幅转为灰度图像
gray_images = [convert_to_gray(block) for block in block_images]

# 3. 对各个图像进行边缘检测
edges_images = [edge_detection(gray) for gray in gray_images]

# 4. 消除噪声干扰
denoised_images = [remove_noise(edges) for edges in edges_images]

# 5. 将各个图幅合并为一个图幅
merged_image = merge_images(denoised_images, block_size)

# 6. 导出结果
export_result(merged_image, "output.jpg")

将边缘检测算子从Canny更改为Sobel的方法是使用cv2.Sobel函数来进行边缘检测。在原代码中的edge_detection函数中，将cv2.Canny替换为cv2.Sobel，并指定相关参数即可。Sobel函数的第一个参数是要进行边缘检测的图像，第二个参数是输出图像的数据类型，第三和第四个参数是指定x和y方向的导数的阶数，第五个参数是指定卷积核的大小。在本例中，我们使用了3x3的卷积核。最后，我们使用cv2.convertScaleAbs函数将结果转换为无符号8位整数类型。