如何平均jpeg图像的通道

I upload an image to r:

我将图片上传到r:

img=readJPEG("18105_552.jpg", native = FALSE)

and the I wrote a for loop to average the pixels:

我写了一个for循环来平均像素:

options(digits=15)
m=matrix(data=NA, nrow=1, ncol=230400)
for(k in 1)
    for (i in (1: 480)){
          for (j in (1: 480)){
           m[k, ]<-((img[i,j,1]+img[i,j,2]+img[i,j,3])/3) 
            }}

However, my output is all ones. If I go in and manually try to get values, they are all .8 or .9 something, so I think r is rounding everything up to 1, but I can't seem to figure out how to stop it from doing so. Any suggestions?

但是,我的输出是全部的。如果我进去并手动尝试获取值,它们都是.8或.9的东西,所以我认为r将所有内容都舍入到1,但我似乎无法弄清楚如何阻止它这样做。有什么建议么?

1 个解决方案

#1

Assuming that img is an array structure or can be converted to one, this can be done without explicit loops using rowMeans():

假设img是一个数组结构或者可以转换为一个,这可以在没有使用rowMeans()的显式循环的情况下完成:

img <- array(runif(30,0.6,1),dim=c(10,10,3))
rowMeans(img,dims=2)

Or, to get it in one row:

或者,将其排成一行:

matrix(rowMeans(img,dims=2),nrow=1)

This stores the numbers columnwise. Use byrow=TRUE to fill up row by row. Next to a correct result, you get a considerable speedup.

这会按列存储数字。使用byrow = TRUE逐行填充。在正确的结果旁边,您可以获得相当快的速度。

By the way, your problem is not rounding. The reason your solution doesn't work, is the following step:

顺便说一句,你的问题不是四舍五入。您的解决方案不起作用的原因如下:

m[k, ]<-((img[i,j,1]+img[i,j,2]+img[i,j,3])/3)

Here you calculate a single number, but assign that to the k'th row of the matrix m. R will recycle the single number, and fill the entire row with that single number. This one number will be the last one you calculated, and that might be the 1 in your case.

在这里计算单个数字,但将其分配给矩阵m的第k行。 R将回收单个数字,并用该单个数字填充整行。这一个数字将是您计算的最后一个数字,在您的情况下可能是1。

EDIT :

To do this for a set of pictures, store them in a list and use sapply to do the work for you:

要对一组图片执行此操作,请将它们存储在列表中并使用sapply为您完成工作:

Say:

img <- array(runif(30,0.6,1),dim=c(10,10,3))
img2 <- array(runif(30,0.6,1),dim=c(10,10,3))

Then make a list:

然后列出一个清单:

images <- list(img,img2)
names(images) <- c('img1','img2')

And :

m <- sapply(images,rowMeans,dims=2)

does what you want.

做你想要的。

#1