Blender has python built in. In the newer versions of Blender, there is even a python console interpreter built right into the scripting section for quick easy debugging and prototyping.
Here is a tutorial with getting started with Blender and Python.
As for actually converting the image, I can't help. You will need to know a lot about 3D data processing and the like. There will be some issues though, because regular 2D images are very hard to grab depth from. You see how long it took Microsoft to develop the Kinect, and it even uses special hardware to grab stereoscopic images of the surroundings.
The Kinect has been open-sourced and has some Python libs to handle the data if that would be a better option.
I'm also looking for a program that does the same thing except I don't care if it's not open source as long as I don't have to pay for it. I found "Autodesk Photo Scene Editor" but it requires you to upload the 20GB of photos to their website then download another 1GB of textures and data to render the model. Surely there is a better way. Please advice.
So there is no way to make such model?
My Real project is described below:
I have some Aerial photograph and am trying to make a model from the mountains.
If you saw before, Geographers make a model from this pictures by Contours.
As I said you can upload 20GB of photos through Autodesk Photo Scene Editor and receive a 1GB 3d model in return but I don't have that sort of bandwidth :( I know it's possible if you can do it through the cloud. The question for me is how to do it without the cloud (eg on localhost).
Somebody please mention a few programs or scripts that will convert a bunch of 2d images to a single textured 3d object (3ds preferably) file. I know Autodesk Photo Scene Editor does this but it requires uploading and downloading each and every photo which takes ever so long. Can somebody please mention an alternative where it is done entirely on my local computer and none of this cloud stuff. (I have an i7 3.0GHz processor) Thank you.