### 使用OpenCV和Canny算法进行工件边缘检测及轮廓特征提取 以下是基于OpenCV库和Canny算法实现工件边缘检测并提取轮廓特征的完整代码示例： ```python import cv2 as cv import numpy as np def edge_detection_and_contour_extraction(image_path): # 读取图像 image = cv.imread(image_path) if image is None: raise ValueError("Image not found or path is incorrect") # 转换为灰度图 gray_image = cv.cvtColor(image, cv.COLOR_BGR2GRAY) # 应用中值模糊减少噪声 blurred_image = cv.medianBlur(gray_image, 5) # 中值滤波[^1] # Canny 边缘检测 edges = cv.Canny(blurred_image, threshold1=50, threshold2=150) # 双阈值设置[^3] # 查找轮廓 contours, hierarchy = cv.findContours(edges, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE) # 绘制轮廓 contour_image = image.copy() cv.drawContours(contour_image, contours, -1, (0, 255, 0), 2) # 提取轮廓特征 features = [] for cnt in contours: area = cv.contourArea(cnt) # 计算面积 perimeter = cv.arcLength(cnt, True) # 计算周长 approx = cv.approxPolyDP(cnt, 0.02 * perimeter, True) # 多边形逼近 num_sides = len(approx) # 获取多边形顶点数量 moments = cv.moments(cnt) # 计算矩 cx = int(moments['m10'] / moments['m00']) # 计算质心x坐标 cy = int(moments['m01'] / moments['m00']) # 计算质心y坐标 feature_dict = { 'area': area, 'perimeter': perimeter, 'num_sides': num_sides, 'centroid': (cx, cy), } features.append(feature_dict) # 显示结果 cv.imshow('Original Image', image) cv.imshow('Edges', edges) cv.imshow('Contours', contour_image) cv.waitKey(0) cv.destroyAllWindows() return features if __name__ == '__main__': input_image_path = '../opencv-python-img/workpiece.jpg' # 替换为实际图片路径 extracted_features = edge_detection_and_contour_extraction(input_image_path) print("Extracted Contour Features:") for idx, feat in enumerate(extracted_features): print(f"Contour {idx + 1}: Area={feat['area']}, Perimeter={feat['perimeter']}, " f"Sides={feat['num_sides']}, Centroid={feat['centroid']}") # 输出轮廓特征 ``` #### 解析 - **中值滤波**：为了降低图像中的噪声影响，在应用Canny算法之前先进行了中值滤波处理[^1]。 - **Canny边缘检测**：通过双阈值参数`threshold1`和`threshold2`调整边缘检测的效果[^3]。较低的阈值用于捕捉弱边缘，较高的阈值则捕获强边缘。 - **查找轮廓**：使用`cv.findContours`函数找到所有的闭合边界，并绘制出来以便可视化。 - **轮廓特征提取**：对于每一个轮廓，计算其面积、周长、多边形顶点数以及质心位置。 --- ###