Abstract—This paper describes an improved Hue-Assisted Normal Distributions Transform (HANDT), which enhances the accuracy of the registration of point clouds. In our previous work, HANDT was developed to improve the speed and accuracy of registration by utilizing the hue from the huesaturation- value model. To improve the accuracy, the score function of HANDT is changed from the normal distribution function to the Mahalanobis distance function. In addition, the functions of the hue mean and variance are modified for the circular property of the hue. The performance of normal distributions transform (NDT), HANDT, and improved HANDT are evaluated by benchmark data sets. As a result, the translation and rotation errors of improved HANDT are lower than those of NDT and HANDT.
Index Terms—Hue, color, point cloud, registration, normal distributions transform (NDT).
The authors are with the Seoul National University, Department of Electrical and Computer Engineering, Seoul, Republic of Korea (e-mail: ckowl@snu.ac.kr, bhlee@snu.ac.kr).
[PDF]
Cite:Hyunki Hong and Beomhee Lee, "Colored Point Cloud Registration with Improved Hue- Assisted Normal Distributions Transform," International Journal of Computer Theory and Engineering vol. 8, no. 1, pp. 63-68, 2016.
