利用哈夫曼编码压缩图像的代码

以下是利用哈夫曼编码压缩图像的 Python 代码示例：

import heapq
import numpy as np

# 定义哈夫曼树节点类
class HuffmanNode:
    def __init__(self, freq, pixel=None):
        self.freq = freq
        self.pixel = pixel
        self.left = None
        self.right = None

    # 定义节点比较方法
    def __lt__(self, other):
        return self.freq < other.freq

# 计算像素值频率
def calc_freq(img):
    freq = {}
    for i in range(img.shape[0]):
        for j in range(img.shape[1]):
            pixel = img[i][j]
            if pixel in freq:
                freq[pixel] += 1
            else:
                freq[pixel] = 1
    return freq

# 构建哈夫曼树
def build_tree(freq):
    heap = []
    for pixel in freq:
        heapq.heappush(heap, HuffmanNode(freq[pixel], pixel))
    while len(heap) > 1:
        node1 = heapq.heappop(heap)
        node2 = heapq.heappop(heap)
        parent = HuffmanNode(node1.freq + node2.freq)
        parent.left = node1
        parent.right = node2
        heapq.heappush(heap, parent)
    return heap[0]

# 递归遍历哈夫曼树，生成编码表
def traverse(node, code, code_table):
    if node.pixel is not None:
        code_table[node.pixel] = code
    else:
        traverse(node.left, code + '0', code_table)
        traverse(node.right, code + '1', code_table)

# 哈夫曼编码压缩图像
def huffman_compress(img):
    freq = calc_freq(img)
    tree = build_tree(freq)
    code_table = {}
    traverse(tree, '', code_table)
    compressed_img = ''
    for i in range(img.shape[0]):
        for j in range(img.shape[1]):
            pixel = img[i][j]
            compressed_img += code_table[pixel]
    # 将压缩后的二进制字符串转换为字节数组
    compressed_bytes = bytearray()
    for i in range(0, len(compressed_img), 8):
        compressed_bytes.append(int(compressed_img[i:i+8], 2))
    return compressed_bytes, code_table

# 哈夫曼解压缩图像
def huffman_decompress(compressed_bytes, code_table, width, height):
    compressed_bits = ''
    for byte in compressed_bytes:
        compressed_bits += bin(byte)[2:].zfill(8)
    img = np.zeros((height, width), dtype=np.uint8)
    node = build_tree({pixel: len(code_table[pixel]) for pixel in code_table})
    for i in range(height):
        for j in range(width):
            node = node.left if compressed_bits[i*width+j] == '0' else node.right
            if node.pixel is not None:
                img[i][j] = node.pixel
                node = build_tree({pixel: len(code_table[pixel]) for pixel in code_table})
    return img

在上述代码中，`calc_freq()` 函数用于计算像素值频率，`build_tree()` 函数用于构建哈夫曼树，`traverse()` 函数用于递归遍历哈夫曼树，生成编码表，`huffman_compress()` 函数用于哈夫曼编码压缩图像，`huffman_decompress()` 函数用于哈夫曼解压缩图像。这里使用了 Python 标准库中的 `heapq` 模块实现优先队列，以便构建哈夫曼树。