写一个找到图片中物体轮廓的曲率突变的点坐标的python程序，调用opencv

在Python中，使用OpenCV库可以方便地处理图像并找出物体轮廓的曲率突变点。下面是一个简单的示例，展示了如何使用`cv2.findContours`函数获取轮廓，并通过霍夫圆检测算法计算曲线的弧度变化，从而定位到突变点：

import cv2
import numpy as np

def find_contour_curve_points(image_path):
    # 读取图像
    image = cv2.imread(image_path, cv2.IMREAD_GRAYSCALE)
    
    # 高斯模糊有助于边缘检测
    blurred_image = cv2.GaussianBlur(image, (5, 5), 0)

    # 转换为二值图像
    _, thresholded = cv2.threshold(blurred_image, 127, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)

    # 寻找轮廓
    contours, _ = cv2.findContours(thresholded, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

    if contours:
        for contour in contours:
            # 计算轮廓的外接矩形
            x, y, w, h = cv2.boundingRect(contour)

            # 提取轮廓像素
            contour_image = thresholded[y:y+h, x:x+w]

            # 使用霍夫圆变换寻找可能的圆心点
            circles = cv2.HoughCircles(contour_image, cv2.HOUGH_GRADIENT, dp=1, minDist=20, param1=50, param2=30, minRadius=0, maxRadius=0)

            # 如果找到圆形，检查相邻像素的弧度差异
            if circles is not None:
                circles = np.round(circles[0, :]).astype("int")
                for (x, y, r) in circles:
                    center = (x + x, y + y)
                    # 计算周围像素的弧度差
                    neighbors = contour_image[max(0, y-r):min(y+r+1, h), max(0, x-r):min(x+r+1, w)]
                    angles = np.arctan2(neighbors[1:] - neighbors[:-1], neighbors[1:, 0] - neighbors[:-1, 0])
                    
                    # 可能的突变点即弧度差显著的位置
                    curvature_derivative = np.abs(np.diff(angles))
                    peak_positions = np.where(curvature_derivative > np.mean(curvature_derivative) * 0.8)[0]
                    peak_positions += x - r  # 将像素坐标转换回原始图像位置
                    
                    yield center, peak_positions
    else:
        print(f"No contours found in the image at {image_path}")

# 使用函数
for point_info in find_contour_curve_points('your_image_path.jpg'):
    print(f"Detected curve change point: ({point_info[0][0]}, {point_info[0][1]}), neighboring peaks: {point_info[1]}")