请给出完整代码仿真实现基于DWT的邻近值水印算法matlmap

好的，以下是基于 Python 语言仿真实现基于DWT的邻近值水印算法matlmap的完整代码：

import cv2
import numpy as np
from math import log2, ceil

def binary_to_decimal(binary):
    """
    将二进制序列转换为十进制数值
    """
    decimal = 0
    for i in range(len(binary)):
        decimal += binary[i] * pow(2, len(binary) - i - 1)
    return decimal

def decimal_to_binary(decimal, length):
    """
    将十进制数值转换为指定长度的二进制序列
    """
    binary = []
    while decimal > 0:
        binary.insert(0, decimal % 2)
        decimal //= 2
    while len(binary) < length:
        binary.insert(0, 0)
    return binary

def matlmap_embed(img_path, watermark, alpha):
    """
    利用邻近值水印算法将水印嵌入到图像中
    img_path: 原始图像路径
    watermark: 水印信息，字符串形式
    alpha: 嵌入强度，取值范围为[0, 1]
    """
    # 加载原始图像
    img = cv2.imread(img_path, cv2.IMREAD_GRAYSCALE)
    rows, cols = img.shape

    # 对图像进行二维离散小波变换，得到四个子带的系数矩阵
    level = ceil(log2(min(rows, cols))) - 1
    coeffs = pywt.wavedec2(img, 'haar', level=level)
    LL, (LH, HL, HH) = coeffs[0], coeffs[1], coeffs[2], coeffs[3]

    # 将水印信息转换为二进制序列
    watermark = ''.join(map(lambda x: '{:08b}'.format(ord(x)), watermark))
    w_length = len(watermark)

    # 计算嵌入强度对应的阈值
    max_HH = np.max(HH)
    threshold = alpha * max_HH

    # 将水印信息嵌入到HH子带的系数矩阵中
    w_index = 0
    for i in range(rows // 2, rows):
        for j in range(cols // 2, cols):
            if abs(HH[i][j]) > threshold:
                HH[i][j] += (-1) ** int(watermark[w_index])
                w_index += 1
                if w_index == w_length:
                    break
        if w_index == w_length:
            break

    # 对修改后的系数矩阵进行离散小波反变换，得到带有水印的图像
    coeffs_new = (LL, (LH, HL, HH))
    img_new = pywt.waverec2(coeffs_new, 'haar')

    return img_new

def matlmap_extract(img_path, alpha, w_length):
    """
    从图像中提取嵌入的水印信息
    img_path: 带有水印的图像路径
    alpha: 嵌入强度，取值范围为[0, 1]
    w_length: 水印信息的长度
    """
    # 加载带有水印的图像
    img = cv2.imread(img_path, cv2.IMREAD_GRAYSCALE)
    rows, cols = img.shape

    # 对图像进行二维离散小波变换，得到四个子带的系数矩阵
    level = ceil(log2(min(rows, cols))) - 1
    coeffs = pywt.wavedec2(img, 'haar', level=level)
    LL, (LH, HL, HH) = coeffs[0], coeffs[1], coeffs[2], coeffs[3]

    # 计算嵌入强度对应的阈值
    max_HH = np.max(HH)
    threshold = alpha * max_HH

    # 从HH子带的系数矩阵中提取水印信息
    watermark = ''
    w_index = 0
    for i in range(rows // 2, rows):
        for j in range(cols // 2, cols):
            if abs(HH[i][j]) > threshold:
                watermark += str(abs(int(HH[i][j])) % 2)
                w_index += 1
                if w_index == w_length:
                    break
        if w_index == w_length:
            break

    # 将二进制序列转换为字符串
    w_chars = []
    for i in range(0, len(watermark), 8):
        w_chars.append(chr(binary_to_decimal(list(map(int, watermark[i:i+8])))))
    watermark_str = ''.join(w_chars)

    return watermark_str

# 测试样例
if __name__ == '__main__':
    img_path = 'lena.png'
    watermark = 'This is a watermark.'
    alpha = 0.1
    img_embed = matlmap_embed(img_path, watermark, alpha)
    cv2.imwrite('lena_watermarked.png', img_embed)
    watermark_extract = matlmap_extract('lena_watermarked.png', alpha, len(watermark))
    print('Extracted watermark:', watermark_extract)

在代码中，matlmap_embed() 函数实现了将水印信息嵌入到图像中的过程，matlmap_extract() 函数实现了从带有水印的图像中提取水印信息的过程。具体实现过程和原理在前面的回答中已经说明，这里不再赘述。