import numpy as np
import pandas as pd
from scipy.spatial import KDTree
from PIL import Image

# 读取CSV文件
data = np.genfromtxt('PHA SID230828-2003.csv', delimiter=',')

# 创建矩阵A
A = np.zeros((120, 120))
A[:data.shape[0], :data.shape[1]] = data

# 创建矩阵B
B = np.zeros((1280, 1024))

# 进行映射和填充已知点
for u in range(120):
    for v in range(120):
        x = 0.211 * u - 4.538 * v + 894.754
        y = -4.594 * u - 0.1044 * v + 800.7384

        # 判断坐标是否在矩阵B的范围内
        if 0 <= int(x) < 1280 and 0 <= int(y) < 1024:
            B[int(x), int(y)] = A[u, v]

# 寻找未知点的最近邻并进行填充
known_points = np.argwhere(B != 0)  # 获取已知点的坐标
tree = KDTree(known_points)  # 构建KD树

unknown_points = np.argwhere(B == 0)  # 获取未知点的坐标
for u, v in unknown_points:
    x = 0.211 * u - 4.538 * v + 894.754
    y = -4.594 * u - 0.1044 * v + 800.7384

    # 寻找最近邻点的索引
    _, idx = tree.query([u, v])
    nearest_point = known_points[idx]
    nearest_u, nearest_v = nearest_point[0], nearest_point[1]

    # 用最近邻点的值填充未知点
    B[u, v] = B[nearest_u, nearest_v]

# 将矩阵B保存为CSV文件
df = pd.DataFrame(B)
df.to_csv('output.csv', index=False, header=False)

matrix_B = B.copy()
# 将矩阵B中的非零值映射到灰度值范围
matrix_B = np.where(matrix_B != 0, matrix_B * 217, matrix_B)

# 将矩阵B转换为8位灰度图像
image = Image.fromarray(matrix_B.astype(np.uint8))

# 保存为BMP图像
image.save('output2.bmp')