Dabae Kim（M2）, Department of Precision Engineering, received 2019 SICE Young Author’s Awards for Practical Application

On 28th February 2020, Dabae Kim (M2), Department of Precision Engineering, received 2019 SICE Young Author’s Awards for Practical Application.

This award is presented to those who have made excellent research reports at conferences or symposiums hosted by the Society of Instrument and Control Engineers (SICE).

<About awarded research>
When operators conduct recovery works at a disaster site, there exists a danger of a secondary disaster. Therefore, it is important to put mobile robots equipped with cameras into the site for working. Especially, in places where human intervention is difficult, it is necessary to perform autonomous environment measurements and self-localization by the mobile robots. At that time, there are many scenes where the mobile robot stops moving and works while rotating its body. For this reason, it is important to estimate the camera rotation. Spherical cameras, which can acquire 360-degree information, are more effective devices for environment measurement than normal cameras. Recently, robust deep learning-based methods to estimate the rotation of the spherical camera have been proposed. However, there exists a non-uniform distortion on equirectangular images obtained from the spherical camera, which may disturb correct learning and accurate estimation. In this research, Equirectangular-Convolutional Neural Network (E-CNN) is proposed to train the network uniformizing the non-uniform distortion and confirmed the accuracy improvement of the rotation estimation.

<Your impression & future plan>
I sincerely grateful to my supervisors for giving me an enthusiastic advice on this research. I also grateful to SICE for appreciating this research and for receiving a great prize. In the future, I will keep my research activities hard.