[Tutor] [tutor] Finding image statistics

Varsha Purohit varsha.purohit at gmail.com
Mon Mar 10 01:53:37 CET 2008


Hello All,
        I had posted this question in this community. I have a trouble.
Actually in my application i need to multiply the resultant value i am
getting with a cellsize. I am getting this cellsize by reading an ascii
file's header. But when i am trying to multiply it with the result of
imagestat i am getting an error.

this is my code.

def ImageStatistics(file1,file2,inFile):

    # find the pixels that are different in two images
    diff = ImageChops.difference(file1,file2)
    #shows the resultant image
    diff.show()
    #convert the different pixels into grayscale so that they are counted
once for each pixel
    diff = ImageOps.grayscale(diff)

 # read the ascii file
    hdr = read_ascii.header(inFile)
    temp= hdr[4].strip().split() # temp is a list which is ['cellsize',
'127']
    cellsize = temp[1]

    def clip(x):
        if x>=1:
            return 1
        else:
            return 0
        # apply the clipping function to eval
    diff = Image.eval(diff,clip)


    #return the resultant area
    val = ImageStat.Stat(diff).sum
    area = val * cellsize
    print "area is ",area
    return area

i am getting an error that i cannot combine imagestat with cellsize and i
cannot multiply it like that. But in my application i need to multiply
cellsize with the value i am getting in the image.eval....

i tried to multiply cellsize with 1 in the clip function where i am writing
return 1*cellsize.

but even it is not working .... can anyone tell me what i am doing wrong
here ???
On Mon, Mar 3, 2008 at 12:55 PM, Kent Johnson <kent37 at tds.net> wrote:

> Varsha Purohit wrote:
> > Yeahh so by doing this i am counting only the difference part since we
> > have grayscaled the image and assuming it will count only the pixels
> > that evolve as difference....
>
> Yes
>
> > if i use sum2 instead of sum i think  it
> > will give squared sum which is area... and if i just use count it would
> > count the number of pixels developed like that... sounds interesting ..
> > thanks for throwing light for me in right direction....
>
> No. First, pixels are already a measure of area. Second, if each pixel
> value is 0 or 1, squaring the values won't make any difference.
>
> Kent
>



-- 
Varsha Purohit,
Graduate Student
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/tutor/attachments/20080309/86823ad5/attachment.htm 


More information about the Tutor mailing list