DIL

Drone Imaging Language

Installation

DIL has a few dependencies. It's written in Python, 3.5+ should work. It also requires Hugin for stitching panoramas. To install the dependencies on Ubuntu:

$ sh setup.sh

It will prompt you for superuser rights. The script is only 5 lines, you can double check it if you like.

Usage

$ python dil.py my_file.dil

DIL Syntax

DIL files are plain text. They have the file extension .dil. Currently 5 commands are supported:

load
highlight
stitch
show
save

DIL files contain 1 command per line. Commands can occur in any order, but certain orderings may be nonsensical. For example, trying to save before loading any images won't do anything.

Command syntax

load

Load images to be processed. Can contain more than one filename. Loading a set of files will clear any previous loads from memory. Unsaved changes will be lost, use save before loading if you want to retain highlights or stitched panoramas.

Example:

load filename_1.jpg filename2.jpg filename3.jpg ...

highlight

Highlight objects in the loaded image set. More than one object class may be listed. All Object detection isn't perfect - results depend on image quality, orientation of the object, and how the CPU is feeling on a given day. Supported object classes can be found in lib/classes.txt.

Example:

highlight car bench person

stitch

Stitch the loaded image set into a 360 panorama. Doesn't take any parameters. If you save after using stitch, the image will be saved as panorama.jpg and an HTML visualizer will be created at index.html. The visualizer lets you zoom and pan around the image.

The raw panorama isn't usually good to look at, it is projected in such a way that 360 pano viewers can un-warp it for human viewing e.g. in a web browser.

Example:

stitch

show

Show the panorama created by stitch. It launches your web browser and displays the visualizer. It doesn't take any parameters.

If you use show without previously stitching, you won't get a panorama :)

Example:

show

save

Save the current files. It takes no parameters. If you have used highlight, any detected objects will be saved in those images. If you have created a panorama with stitch, the raw panorama image will be saved to panorama.jpg and the visualizer demo will be saved to index.html.

Future possibilities

• file globbing
• maybe ocr
• ???