import cv2
import numpy as np
import matplotlib.pyplot as plt
import json

# 读取图片
img_path = r"C:\Users\yangy\.cursor\projects\h-soft-projects-luxsin-dashboard/assets/c__Users_yangy_AppData_Roaming_Cursor_User_workspaceStorage_b134a9df77916b35c1e5b1ece8dc14fe_images_Arcona-avg-0b78da44-4aaa-465b-9b2b-fba051a28742.png"
img = cv2.imread(img_path)
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img_copy = img.copy()
gray = cv2.cvtColor(img_copy, cv2.COLOR_RGB2GRAY)

print(f"图片尺寸：{img_copy.shape}")
height, width = gray.shape

# 定义图表区域（排除顶部标题和底部图例）
# 根据图片估算：顶部约 5%，底部约 8%
chart_top = int(height * 0.05)
chart_bottom = int(height * 0.92)
chart_left = int(width * 0.02)
chart_right = int(width * 0.98)

# 裁剪出图表区域
chart_roi = gray[chart_top:chart_bottom, chart_left:chart_right]

# 使用阈值提取白色线条
_, thresh = cv2.threshold(chart_roi, 200, 255, cv2.THRESH_BINARY)

# 形态学操作，连接断开的线条
kernel = np.ones((3,3), np.uint8)
dilated_thresh = cv2.dilate(thresh, kernel, iterations=2)
eroded_thresh = cv2.erode(dilated_thresh, kernel, iterations=1)

# 查找轮廓
contours, _ = cv2.findContours(eroded_thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

# 找到最大的连续轮廓（应该是频响曲线）
largest_contour = None
max_area = 0

for contour in contours:
    area = cv2.contourArea(contour)
    if area > max_area:
        max_area = area
        largest_contour = contour

print(f"最大轮廓面积：{max_area}")

# 提取频响曲线上的点
frequency_points = []

if largest_contour is not None:
    # 对于每个 x 坐标，找到对应的 y 坐标（取平均值）
    x_coords = {}
    
    for point in largest_contour:
        x, y = point[0]
        # 转换回原图坐标
        global_x = int(x) + chart_left
        global_y = int(y) + chart_top
        
        if global_x not in x_coords:
            x_coords[global_x] = []
        x_coords[global_x].append(global_y)
    
    # 对每个 x，计算平均 y 值
    for x in sorted(x_coords.keys()):
        y_values = x_coords[x]
        avg_y = sum(y_values) // len(y_values)
        frequency_points.append([int(x), int(avg_y)])
    
    print(f"提取的频响曲线点数：{len(frequency_points)}")
    
    # 创建可视化结果
    result = img_copy.copy()
    
    # 绘制检测到的点
    for i, (x, y) in enumerate(frequency_points):
        cv2.circle(result, (x, y), 1, (255, 0, 0), -1)
    
    # 显示结果
    plt.figure(figsize=(20, 10))
    plt.imshow(result)
    plt.title(f'Extracted Frequency Response Curve ({len(frequency_points)} points)')
    plt.axis('off')
    plt.tight_layout()
    plt.show()
    
    # 保存点到 JSON 文件
    output_data = {
        "frequency_points": frequency_points,
        "chart_area": {
            "top": int(chart_top),
            "bottom": int(chart_bottom),
            "left": int(chart_left),
            "right": int(chart_right)
        },
        "image_size": {
            "width": int(width),
            "height": int(height)
        }
    }
    
    with open('frequency_response_points.json', 'w', encoding='utf-8') as f:
        json.dump(output_data, f, ensure_ascii=False, indent=2)
    
    print(f"\n频响点位已保存到 frequency_response_points.json")
    print(f"\n前 50 个点的坐标 (x, y):")
    for i, point in enumerate(frequency_points[:50]):
        print(f"{i+1}: {point}")
    
    # 也保存为 CSV 格式，方便查看
    with open('frequency_response_points.csv', 'w', encoding='utf-8') as f:
        f.write("index,x,y\n")
        for i, point in enumerate(frequency_points):
            f.write(f"{i+1},{point[0]},{point[1]}\n")
    
    print(f"\n点位也已保存到 frequency_response_points.csv")
else:
    print("未找到频响曲线")