Joint temperature estimation and nonuniformity correction in low-cost infrared cameras - סמינר מחלקה פיסיקלית

סמינר זה יחשב כסמינר שמיעה לתלמידי תואר שני—-

You are invited to attend a lecture on Thursday, Jul. 13, 2023 ,15:00

Room 011, Kitot Building

Zoom https://tau-ac-il.zoom.us/j/87605414407

Joint temperature estimation and nonuniformity correction in low-cost infrared cameras

By:

Navot Oz

PhD student under the supervision of Prof. David Mendelovic, Prof. Nir Sochen and Dr. Iftach Klapp

Abstract

Infrared (IR) cameras are used for temperature measurements in various applications, but low-cost microbolometer-based IR cameras suffer from inaccurate and spatially variant nonuniformity, which depends on the ambient temperature of the camera. These limitations reduce the usability and reliability of low-cost IR cameras compared to expensive radiometric cameras.

In this work, we propose two novel methods for improving the temperature accuracy and correcting the nonuniformity of low-cost IR cameras, using only the ambient temperature measured by the camera itself. The first method uses a neural network based on the physical model of the camera that estimates the object's temperature and rectifies the nonuniformity from a single image. The second method is based on a kernel estimation network that combines multiple frames captured by the camera, despite imperfect registration, using a physical image acquisition model and accounts for the ambient temperature.

Both methods achieve significant improvement in performance compared to previous work. The methods were evaluated on real data collected by a low-cost IR camera mounted on a UAV, showing accuracy like scientific radiometric cameras.