Remote Sensing of Gases by Hyperspectral Imaging

Idea behind the system is to detect gases through a remote device using infrared (IR) camera. Data of the surrounding is collected, and then image data is converted to frequency domain to compare it with the spectrum of the gas (e.g., Ammonia, Methanol, SF6, etc.). A network of the remote sensing devices is shown in the following Figure.

A Network for Remote Sensing of Gases by Hyperspectral Imaging

Software

I have developed software for hyperspectral imaging. Following hardwares are controlled using the software.

  • Voice-coil motor with controller (DMC 30000)
  • Focal plane arrays (FPA), Infrared Camera with Megalink electronic board through camera link
  • Image processing of CCD camera and IR camera
  • Frame grabber (camera link) to receive pixel data
  • Fast steering mirror (FSM)
  • Pan/Tilt control (Moog's product)
  • Serial & Ethernet Communication

Software includes graphs of interferogram and FFT spectrum, nonuniformity correction algorithm, bad pixel correction algorithm, and other algorithms. I used FLTK library for GUI, OpenCV library to show graphs and for image processing, FFTW library to perform FFT, and other libraries to control hardware. Furthermore, I developed a UDP-based communication protocol to send and receive data between computer and DSP. Overview of the software is represented in following Figures.

Software for Remote Sensing of Gases by Hyperspectral Imaging
Software for Remote Sensing of Gases by Hyperspectral Imaging

Nonuniformity Correction of IR camera

I have developed an algorithm of nonuniformity correction for IR camera. The proposed algorithm is based on two-point callibration. I used two reference images at 30-degree and 40-degree temperatures. The reference images were captured by placing IR camera infront of Blackbody. In the following Figure, both images, before and after nonunifomity correction, are shown.

IR image of Irfan Ullah before nonuniformity correction IR image after nonuniformity correction

Hardware

For hardware implementation, we used IR camera of 320x256 pixels, as shown in the following Figure.

Inrfrared (IR) Camera

To get spectrum from IR camera, we used voice coil motor to move the FSM, as shown in the following Figure.

Voice Coil Motor