1
2
3
4

# 更新pip
python -m pip install -U pip
# 安装 scikit-image
python -m pip install -U scikit-image

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113

import cv2
import math
import os
import numpy as np
import matplotlib.pyplot as plt
from skimage.metrics import mean_squared_error as compare_mse
from skimage.metrics import peak_signal_noise_ratio as compare_psnr
from skimage.metrics import structural_similarity as compare_ssim
from skimage.measure import shannon_entropy as compare_entropy

def compare_img(img_raw_path, img_fusion_path):
    ssim_list = []
    rmse_list = []
    psnr_list = []
    en_list = []
    ag_list = []
    for filename_raw ,filename_fusion in zip(os.listdir(img_raw_path),os.listdir(img_fusion_path)):
        img_raw = cv2.imread(img_raw_path+'/'+filename_raw)
        img_fusion = cv2.imread(img_fusion_path+'/'+filename_fusion)
        ssim = compare_ssim(img_raw, img_fusion, channel_axis=2) # 对于多通道图像(RGB、HSV等)关键词multichannel要设置为True
        rmse = math.sqrt(compare_mse(img_raw, img_fusion))
        psnr = compare_psnr(img_raw, img_fusion)
        en = compare_entropy(img_fusion)
        ag = compare_ag(img_fusion)
        ssim_list.append(ssim)
        rmse_list.append(rmse)
        psnr_list.append(psnr)
        en_list.append(en)
        ag_list.append(ag)        
    return ssim_list, rmse_list, psnr_list,en_list,ag_list

def compare_ag(image):
    image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    image = np.array(image)
    image = image.astype(np.float32)
    width = image.shape[1]
    width = width - 1
    height = image.shape[0]
    height = height - 1
    tmp = 0
    for i in range(1,height):
        for j in range(1,width):
            dx = image[i][j + 1] - image[i][j]
            dy = image[i + 1][j] - image[i][j]
            ds = math.sqrt((dx * dx + dy * dy) / 2)
            tmp += ds
    return tmp / (height * width)

def cal_mean_metric(list1,list2):
    average_metric = [list1,list2]
    average_metric = list(np.mean(average_metric, axis=0))
    return average_metric

def plot_metric(list1,list2,name):
    plt.figure()
    img_nums = len(list1)
    x_axis_data = range(1, img_nums+1)
    plt.plot(x_axis_data, list1, 'b*--',label="Original method "+name ,linewidth=1)
    plt.plot(x_axis_data, list2, 'r.--',label="Improved method "+name,linewidth=1)
    plt.legend()
    plt.xticks(x_axis_data)
    plt.xlabel("Image pair", fontsize=13, fontweight='bold')
    plt.ylabel("Metrics Value", fontsize=13, fontweight='bold')


if __name__ == "__main__":
    img_rgb_path = './img/rgb'
    img_red_path = './img/red'
    img_fusion_path = './img/img_fusion/'
    img_improve_path = './img/img_improve_fusion'

    rgb_ssim_list,rgb_rmse_list,rgb_psnr_list,rgb_en_list,rgb_ag_list = compare_img(img_rgb_path, img_fusion_path)
    red_ssim_list,red_rmse_list,red_psnr_list,red_en_list,red_ag_list = compare_img(img_red_path, img_fusion_path)
    average_ssim = cal_mean_metric(rgb_ssim_list,red_ssim_list)
    average_rmse = cal_mean_metric(rgb_rmse_list,red_rmse_list) 
    average_psnr = cal_mean_metric(rgb_psnr_list,red_psnr_list)
    # for i in range(len(average_ssim)):
    #     print("改进前:Average SSIM: %.4f" %average_ssim[i],
    #           "Average RMSE: %.4f" %average_rmse[i],
    #           "Average PSNR: %.4f" %average_psnr[i],
    #           "EN: %.4f" %rgb_en_list[i],
    #           "AG: %.4f" %rgb_ag_list[i])

    print("改进前:Average SSIM: %.4f" %np.mean(rgb_ssim_list),
            "Average RMSE: %.4f" %np.mean(rgb_rmse_list),
            "Average PSNR: %.4f" %np.mean(rgb_psnr_list),
            "EN: %.4f" %np.mean(rgb_en_list),
            "AG: %.4f" %np.mean(rgb_ag_list))

    rgb_ssim_list1,rgb_rmse_list1,rgb_psnr_list1,rgb_en_list1,rgb_ag_list1 = compare_img(img_rgb_path, img_improve_path)
    red_ssim_list1,red_rmse_list1,red_psnr_list1,red_en_list1,red_ag_list1 = compare_img(img_red_path, img_improve_path)
    average_ssim1 = cal_mean_metric(rgb_ssim_list1,red_ssim_list1)
    average_rmse1 = cal_mean_metric(rgb_rmse_list1,red_rmse_list1) 
    average_psnr1 = cal_mean_metric(rgb_psnr_list1,red_psnr_list1)
    # for i in range(len(average_ssim1)):
    #     print("改进后:Average SSIM: %.4f" %average_ssim1[i],
    #           "Average RMSE: %.4f" %average_rmse1[i],
    #           "Average PSNR: %.4f" %average_psnr1[i],
    #           "EN: %.4f" %rgb_en_list1[i],
    #           "AG: %.4f" %rgb_ag_list1[i])

    print("改进后:Average SSIM: %.4f" %np.mean(rgb_ssim_list1),
            "Average RMSE: %.4f" %np.mean(rgb_rmse_list1),
            "Average PSNR: %.4f" %np.mean(rgb_psnr_list1),
            "EN: %.4f" %np.mean(rgb_en_list1),
            "AG: %.4f" %np.mean(rgb_ag_list1))

    plot_metric(average_ssim,average_ssim1,"SSIM")
    plot_metric(average_rmse,average_rmse1,"RMSE")
    plot_metric(average_psnr,average_psnr1,"PSNR")
    plot_metric(rgb_en_list,rgb_en_list1,"EN")
    plot_metric(rgb_ag_list,rgb_ag_list1,"AG")
    plt.show()