canny边缘检测算子实现边缘检测

Canny边缘检测是一种常用的边缘检测算法，其主要步骤包括：

1. 高斯滤波：使用高斯滤波器对图像进行平滑处理，以去除噪声。

2. 计算梯度幅值和方向：对平滑后的图像计算每个像素点的梯度幅值和方向。

3. 非极大值抑制：对梯度图像进行非极大值抑制，以保留边缘的细节。

4. 双阈值处理：将梯度图像按照两个阈值进行二值化处理，较大的梯度幅值被认为是边缘，较小的梯度幅值被认为是噪声或非边缘。

5. 连通区域分析：对二值化后的图像进行连通区域分析，以得到完整的边缘。

下面是Canny边缘检测算法的Python实现：

import cv2
import numpy as np

def canny_edge_detection(img, low_threshold, high_threshold):
    # 高斯滤波
    img = cv2.GaussianBlur(img, (3,3), 0)
    
    # 计算梯度幅值和方向
    dx = cv2.Sobel(img, cv2.CV_16S, 1, 0)
    dy = cv2.Sobel(img, cv2.CV_16S, 0, 1)
    gradient = np.sqrt(dx*dx + dy*dy)
    gradient_direction = np.arctan2(dy, dx) * 180 / np.pi
    
    # 非极大值抑制
    rows, cols = img.shape
    for i in range(1, rows-1):
        for j in range(1, cols-1):
            if (gradient_direction[i,j] <= 22.5 and gradient_direction[i,j] > -22.5) or \
               (gradient_direction[i,j] <= -157.5 and gradient_direction[i,j] > 157.5):
                if gradient[i,j] > gradient[i,j+1] and gradient[i,j] > gradient[i,j-1]:
                    gradient[i,j] = gradient[i,j]
                else:
                    gradient[i,j] = 0
            elif (gradient_direction[i,j] <= 67.5 and gradient_direction[i,j] > 22.5) or \
                 (gradient_direction[i,j] <= -112.5 and gradient_direction[i,j] > -157.5):
                if gradient[i,j] > gradient[i-1,j+1] and gradient[i,j] > gradient[i+1,j-1]:
                    gradient[i,j] = gradient[i,j]
                else:
                    gradient[i,j] = 0
            elif (gradient_direction[i,j] <= 112.5 and gradient_direction[i,j] > 67.5) or \
                 (gradient_direction[i,j] <= -67.5 and gradient_direction[i,j] > -112.5):
                if gradient[i,j] > gradient[i-1,j] and gradient[i,j] > gradient[i+1,j]:
                    gradient[i,j] = gradient[i,j]
                else:
                    gradient[i,j] = 0
            elif (gradient_direction[i,j] <= 157.5 and gradient_direction[i,j] > 112.5) or \
                 (gradient_direction[i,j] <= -22.5 and gradient_direction[i,j] > -67.5):
                if gradient[i,j] > gradient[i-1,j-1] and gradient[i,j] > gradient[i+1,j+1]:
                    gradient[i,j] = gradient[i,j]
                else:
                    gradient[i,j] = 0
    
    # 双阈值处理
    edge_img = np.zeros((rows, cols), dtype=np.uint8)
    strong_edge_row, strong_edge_col = np.where(gradient > high_threshold)
    weak_edge_row, weak_edge_col = np.where((gradient >= low_threshold) & (gradient <= high_threshold))
    edge_img[strong_edge_row, strong_edge_col] = 255
    edge_img[weak_edge_row, weak_edge_col] = 50
    
    # 连通区域分析
    for i in range(1, rows-1):
        for j in range(1, cols-1):
            if edge_img[i,j] == 50:
                if (edge_img[i-1,j-1] == 255) or (edge_img[i-1,j] == 255) or (edge_img[i-1,j+1] == 255) or \
                   (edge_img[i,j-1] == 255) or (edge_img[i,j+1] == 255) or \
                   (edge_img[i+1,j-1] == 255) or (edge_img[i+1,j] == 255) or (edge_img[i+1,j+1] == 255):
                    edge_img[i,j] = 255
                else:
                    edge_img[i,j] = 0
                    
    return edge_img

其中，`img`为输入的灰度图像，`low_threshold`和`high_threshold`为两个阈值，可以通过调整这两个值来控制边缘检测的效果。函数的返回值为二值化后的边缘图像。