This is the question: can be Blender a powerful tool for computer vision with OpenCv.

Blender is now a powerful suite of 3d content creation, where have a various characteristich as animation, composit, shade, model, texture, interactive ... And OpenCV is ga powerful realtime library for computer Vision.

Then my goal is provide a new and powerfull tool based on Blender for Computer Vision Tasks with OpenCV Library.

How?, with a particular characteristic of Blender, the compositor node.

Nodes are a good tool that provide a user interactive, each node is a box with inputs and outputs that implement a function. Each node have interactive parameters that user can be set.

Each node can be linked with others nodes linking outputs to inputs

In this page we can see more about node compositor

With this properties we can create a workflow of a computer vision tasks.

For example here are the basic edge detection with blenderocv, this example implement the basic opencv sample "edge.c"

Blenderocv - edge (canny)

Now there are a demo with some few functions.

I have to implement a lot of functions as nodes, and now there are a lot of work to do, and sure i make important decisions in development. If anybody want colaborate with this project, please send me mail (david at artresnet dot com) or send me a comment.

There are more info and progress about this project in