Untitled

For people who have been programming in C++, switching to C# in quite smooth because of syntactical similarities. But there are some things to be careful about when you make this switch. One of those things is ‘how the member variables of a class are initialized when there is no explicit default constructor”.

When there is no default constructor for a class then compiler provides it and…

The .ply format is a very common 3D data representation format. This is very widely used because of some of its advantages over other formats-

1. Very easy and simple to understand and interpret.

2. Both Binary and ASCII versions exist, unlike .obj format which has only ASCII version.

3. User defined like XML.

Let me elaborate these points in a bit more detail:

Very easy and simple to understand…

Programmable OpenGL pipeline C++ Skeleton Program Structure

Since i had a hard time to shift from conventional and old fixed openGL programming style to modern OpenGL pipeline which requires us to write shaders. The fixed pipeline is very inefficient and rigid, doesn’t offer much flexibility to graphic programmers. Whereas on the other hand, we have new OpenGL programmable pipeline in which every thing can be controlled by the programmers using GLSL…

Point Cloud Rendering Engine

Point cloud has become a very common 3D data representation format. It has many advantages to it as compared to its mesh counterpart, one of them being low memory needed to store them.

With so much amazing research going on in Computer Vision, it is now possible to generate the 3D point clouds representing the entire cities ( yes!! the entire city literally. Don’t wanna believe me?? check thisout…

Working With Kinect Windows SDK

In this post I want to share few basic sample applications to play with Kinect sensor. I have used microsoft windows SDK as API to work with kinect, OpenGL as graphics rendering API and C++ on Visual Studios 2012 as programming platform.

I will show you 3 different applications namely-

To visualize and capture a colored Point Cloud as .ply file using Kinect.

Given a Kinect Depth Image and a…

Flow Chart of Graphics Rendering Pipeline

The most tricky part of the Rendering Pipeline for me, at-least,  was to understand the concept of fragment.

Fragment: When every primitive in the scene that survives the clipping, backface culling is rasterized we get fragments for each primitive. Thus a fragment is 3D pixel, having following attributes:

position <x,y,z> : x,y is position on the 2D raster grid, and z is the depth of the fragment

Rendering a Colored Point Cloud in BLENDER.

This Image shown above is a Point Cloud Rendered using BLENDER.

The problem with rendering a point cloud in Blender is that it is built to work with meshes. Any mesh format be it .ply, .obj etc has face information . But Point Cloud has only information of vertices and their attributes.  The  point cloud .ply file that I have rendered using the method described in this post has following vertex…

Transformation Matrix used in ‘Kinect’ for Windows to transform depth image to point cloud.

Kinect is am imaging device that also captures the depth of the scene along with its red, green and blue color components.

The depth image in Kinect is expressed as <i,j,d> :

i,j : each pair i,j represent the location of the pixel in the captured frame.

Eg. <280,140> = pixel located at 280 pixels on x-axis and 140 pixels on y-axis.

d : depth at that pixel location expressed in millimeters (mm).