1. 基本原理

在灰度图中，像素值的范围为[0, 255]，即共有256级灰度。在计算机中，我们使用8比特数来表示每一个像素值。因此可以提取出不同比特层面的灰度图。比特层面分层可用于图片压缩：只储存较高比特层（为什么使用较高层，而不是较低层？通过二进制转换，我们知道较高层在数值中的贡献更大）；如使用高四位比特层表示原有的八层比特平面。

2. 测试结果

图源自skimage

3. 代码

 1 def extract_bit_layer(input_image, layer_num):
 2     '''
 3     提取比特层
 4     :param input_image: 原图像
 5     :param layer_num: 提取层
 6     :return: 提取到的比特层
 7     '''
 8     input_image_cp = np.copy(input_image)  # 输入图片的副本
 9 
10     if layer_num == 1:
11         input_image_cp = np.where((input_image_cp >= 0) & (input_image_cp < 2), 255, 0)
12     elif layer_num == 2:
13         input_image_cp = np.where((input_image_cp >= 2) & (input_image_cp < 4), 255, 0)
14     elif layer_num == 3:
15         input_image_cp = np.where((input_image_cp >= 4) & (input_image_cp < 8), 255, 0)
16     elif layer_num == 4:
17         input_image_cp = np.where((input_image_cp >= 8) & (input_image_cp < 16), 255, 0)
18     elif layer_num == 5:
19         input_image_cp = np.where((input_image_cp >= 16) & (input_image_cp < 32), 255, 0)
20     elif layer_num == 6:
21         input_image_cp = np.where((input_image_cp >= 32) & (input_image_cp < 64), 255, 0)
22     elif layer_num == 7:
23         input_image_cp = np.where((input_image_cp >= 64) & (input_image_cp < 128), 255, 0)
24     elif layer_num == 8:
25         input_image_cp = np.where((input_image_cp >= 128) & (input_image_cp < 256), 255, 0)
26     else:
27         print("please enter the number of bit layers from 1 to 8")
28 
29     output_image = input_image_cp
30 
31     return output_image

标签：layer,比特,灰度变换,image,Python,num,input,cp,255
来源： https://www.cnblogs.com/iwuqing/p/11297280.html