Hi

I have a bunch of jpeg images.

say 001abcd.jpg
002abcd.jpg

and so on .

what i want is capture the filename and add it as a text in the image itself in one corner.
so the result would be .. for example the file 003abcd.jpg will have "003abcd" imprinted in one corner of that image. (extension need not be there)


i want a terminal command that can batch process hundreds of images and add its own filenames in its respective images.

i am using linux mint 17

something tells me that imagemagick can be useful..but i dont know scripting.

it is easy to put a single common text in all images. but i dont know how to put unique filenames as the text in the respective images in one go.

Hello, and welcome to the forum. I'm "bumping" this because that sounds like a script I would also like to see. As you say it is easy to rename and add "common text" to all the images.

You are correct, and imagemagick can be used for this. Read their documentation:
http://www.imagemagick.org/Usage/annotating/#annotating

There are MANY easily-found bash scripting tutorials, so start there. We will be happy to HELP you if you're stuck, but this isn't the place to come and have someone write a script for you. Post what YOU have written/tried of your own, and we can help from there.

Otherwise, try looking up "bash script to return file names" in Google, which pulls up lots of methods to do this. Then, look up how to loop through all the files in a directory, and apply the action you would have found from the imagemagick page linked to you above.

You have all the tools needed to do this yourself.

1 members found this post helpful.

There is also graphicsmagick a fork of imagemagick.

Here's a simple for loop script

1. The *.[Jj][Pp][Gg] will cover all image files with a mix of upper and lower case jpg letters.
   
2. The "${i%.*}" will print the filename into the image minus the extension.
   
3. The -gravity East will print the text in the lower right corner

Since you said, you wanted the filenames to match or reference the text in the images, then you will need to export the final results into a different location. The convert command will overwrite files without asking if present in the same folder.

The $HOME/tmp/"$i" does this. Create a tmp folder in the root directory of your home folder.

Note: I am using the tmp folder in this example, but you can use any folder for the final results, just make sure you put the correct path for the newly created images in the code above.

PS: Execute the script on a few test images to see if the results are what you're looking for.

1 members found this post helpful.

I'm not the OP but it works fine for me, thanks. I already had a couple of perl scripts that I was using to batch process (annotate, re-size and rename) my photos. With the genesis of this thread I found and added a "mogrify" statement to the script that does what I wanted, gives me the option to add the file name as a second annotation but it's always nice to see another approach.

Thanks mjolnir.

BTW, I noticed the code you quoted back to me is missing the percent sign '%' in the code.

Not sure if it's firefox or the the LQ forum software. I've noticed that certain characters will not show at all when using the code tag. When I posted the code yesterday, the % sign didn't appeared in the code as well. I was stomp as to why. Then I tried the Go Advanced edit feature and the percent sign appeared in the code as it should be.

It's very weird...

[QUOTE] You solved my case Buffer Overflow and i made few alternatives using your idea.


this command
adds filename over the image as text
in the bottom center
background black
and text is white
the result is all copied in a tmp folder in the home folder



this command

background transparent
and text is white

rest all same


this command

bottom right
background transparent
and text is white
font is dejavusans bold
with black outline
rest all same


thats all
thanks again everyone..

you need not read further i am just describing why i needed this idea.

Infrared Motion detection guide.
the purpose of my question was a project i am trying.
i needed to install a motion detection infrared camera (something like paranormal activity)(i am serious)

i needed
an infrared camera . ( i am going to convert a cheap webcam to infrared) (not done yet)

a motion detection software. ( linux has a software called motion)
it clicks pictures continuously the moment it sees a movement. in the end i get a large number of jpeg images .
a 30 sec movement caused approximately 800 images to be captured by motion software. each jpeg image had the date and time signature as the filename.

motion is working well as expected in my laptop .
next
i have to stitch all the images into a single video and see the result as a continuous video.
this is also successful by using a terminal command

so my whole project is successful as long as im trying it during the day time using a normal webcam.

technically my project is complete for a day camera job.

one question came into my mind.

consider i kept the camera on for 8 hours. and the images got captured whenever there was movement.
i might end up with several thousands of images.

i would even combine the entire images into video.
now while seeing the video
how can one find out at what time did a given movement happened.

for example consider a bird came and sat on the table at 1.30 pm.
definetly the motion will be captured. and i would be able to make the video later.

but how can i find out at what time the bird come on the table.
one difficult way is to see every image in the entire batch of thousands of images and see when the picture of bird was captured. the filename will give the answer for my time.

another idea i thought was to keep a digital table clock in front of the camera view somewhere in the room.
but that was silly. idea.

because everytime the time would change the software would detect motion and would capture unnecessary images. in the end it may capture the entire 8 hours length of images.

no use.

finally i thought . there should be a software technique to imprint the filename onto the image as text somewhere in the bottom. because the filename has the time signature already.

so when i run the video later i ll get to see the time signature also . that will help me to know what time the event occured.

this is exactly what you solved. thanks ..again Buffer Overflow.

so now not only this is a solution to my question . this is also a guide to make a motion detection video.

You''re welcome emmfranklin and welcome to LQ!

@ emmfranklin

I was experimenting with the gravity option. You probably may know this. But you can put the texts in the extreme corners by using NorthEast, NorthWest,SouthWest and SouthEast with the gravity switch.

Attached Thumbnails

       

1 members found this post helpful.

Yes Buffer overflow

You are right again

The SouthEast argument brings the text to the right hand side corner bottom

While East brings it to right centre

All interested readers may experiment with North, East, West, South, or NorthEast,NorthWest, SouthEast, Southwest.