Python計(jì)算圖片數(shù)據(jù)集的均值方差示例詳解

第Python計(jì)算圖片數(shù)據(jù)集的均值方差示例詳解目錄前言Python批量reshape圖片參考計(jì)算數(shù)據(jù)集均值和方差

前言

在做圖像處理的時(shí)候，有時(shí)候需要得到整個(gè)數(shù)據(jù)集的均值方差數(shù)值，以下代碼可以解決你的煩惱：

（做這個(gè)之前一定保證所有的圖片都是統(tǒng)一尺寸，不然算出來(lái)不對(duì)，我的代碼里設(shè)計(jì)的是512*512，可以自己調(diào)整，同一尺寸的代碼我也有：

Python批量reshape圖片

#-*-coding:utf-8-*-

CreatedonThuAug2316:06:352025

@author:libo

fromPILimportImage

importos

defimage_resize(image_path,new_path):#統(tǒng)一圖片尺寸

print('============修改圖片尺寸')

forimg_nameinos.listdir(image_path):

img_path=image_path+"/"+img_name#獲取該圖片全稱(chēng)

image=Image.open(img_path)#打開(kāi)特定一張圖片

image=image.resize((512,512))#設(shè)置需要轉(zhuǎn)換的圖片大小

#processthe1channelimage

image.save(new_path+'/'+img_name)

print("endtheprocessing!")

if__name__=='__main__':

print("readyfor::::::::")

ori_path=r"Z:\pycharm_projects\ssd\VOC2007\JPEGImages"#輸入圖片的文件夾路徑

new_path='Z:/pycharm_projects/ssd/VOC2007/reshape'#resize之后的文件夾路徑

image_resize(ori_path,new_path)

importos

fromPILimportImage

importmatplotlib.pyplotasplt

importnumpyasnp

fromscipy.miscimportimread

filepath=r'Z:\pycharm_projects\ssd\VOC2007\reshape'#數(shù)據(jù)集目錄

pathDir=os.listdir(filepath)

R_channel=0

G_channel=0

B_channel=0

foridxinrange(len(pathDir)):

filename=pathDir[idx]

img=imread(os.path.join(filepath,filename))/255.0

R_channel=R_channel+np.sum(img[:,:,0])

G_channel=G_channel+np.sum(img[:,:,1])

B_channel=B_channel+np.sum(img[:,:,2])

num=len(pathDir)*512*512#這里（512,512）是每幅圖片的大小，所有圖片尺寸都一樣

R_mean=R_channel/num

G_mean=G_channel/num

B_mean=B_channel/num

R_channel=0

G_channel=0

B_channel=0

foridxinrange(len(pathDir)):

filename=pathDir[idx]

img=imread(os.path.join(filepath,filename))/255.0

R_channel=R_channel+np.sum((img[:,:,0]-R_mean)**2)

G_channel=G_channel+np.sum((img[:,:,1]-G_mean)**2)

B_channel=B_channel+np.sum((img[:,:,2]-B_mean)**2)

R_var=np.sqrt(R_channel/num)

G_var=np.sqrt(G_channel/num)

B_var=np.sqrt(B_channel/num)

print("R_meanis%f,G_meanis%f,B_meanis%f"%(R_mean,G_mean,B_mean))

print("R_varis%f,G_varis%f,B_varis%f"%(R_var,G_var,B_var))

可能有點(diǎn)慢，慢慢等著就行。。。。。。。

最后得到的結(jié)果是介個(gè)

參考

計(jì)算數(shù)據(jù)集均值和方差

importos

fromPILimportImage

importmatplotlib.pyplotasplt

importnumpyasnp

fromscipy.miscimportimread

filepath=‘/home/JPEGImages‘#數(shù)據(jù)集目錄

pathDir=os.listdir(filepath)

R_channel=0

G_channel=0

B_channel=0

foridxinxrange(len(pathDir)):

filename=pathDir[idx]

img=imread(os.path.join(filepath,filename))

R_channel=R_channel+np.sum(img[:,:,0])

G_channel=G_channel+np.sum(img[:,:,1])

B_channel=B_channel+np.sum(img[:,:,2])

num=len(pathDir)*384*512#這里（384,512）是每幅圖片的大小，所有圖片尺寸都一樣

R_mean=R_channel/num

G_mean=G_channel/num

B_mean=B_channel/num

R_channel=0

G_channel=0

B_channel=0

foridxinxrange(len(pathDir)):

filename=pathDir[idx]

img=imread(os.path.join(filepath,filename))

R_channel=R_channel+np.sum((img[:,:,0]-R_mean)**2)

G_channel=G_channel+np.sum((img[:,:,1]-G_mean)**2)

B_channel=B_channel+np.sum((img[:,:,2]-B_mean)**2)

R_var=R_channel/num

G_var=G_channel/num

B_var=B_channel/num

print("R_meanis%f,G_meanis%f,B_meani