*Review* **6D Pose Estimation of Objects: Recent Technologies and Challenges**

**Zaixing He 1, Wuxi Feng 1, Xinyue Zhao 1,\* and Yongfeng Lv <sup>2</sup>**


**Abstract:** 6D pose estimation is a common and important task in industry. Obtaining the 6D pose of objects is the basis for many other functions such as bin picking, autopilot, etc. Therefore, many corresponding studies have been made in order to improve the accuracy and enlarge the range of application of various approaches. After several years of development, the methods of 6D pose estimation have been enriched and improved. Although some predecessors have analyzed the methods and summarized them in detailed, there have been many new breakthroughs in recent years. To understand 6D pose estimation better, this paper will make a new and more detailed review of 6D pose estimation. We divided these methods into two approaches: Learning-based approaches and non-learning-based approaches, including 2D-information-based approach and 3D-informationbased approach. Additionally, we introduce the challenges that exist in 6D pose estimation. Finally, we compare the performance of different methods qualitatively and discuss the future development trends of the 6D pose estimation.

**Keywords:** 6D pose estimation; learning-based approach; 2D-information-based approach; 3D-information-based approach; textureless and reflective objects; foreground occlusion; background clutter

**1. Introduction**

#### *1.1. Overview*

6D pose refers to the posture of an object, specifically on the basis of a translation vector and a rotation vector. 6D pose estimation is an important step in many industrial fields highly related to another challenge—problem tracking [1]—such as bin picking [2–6], autonomous driving [7–9], augmented reality [10–12], SLAM (Simultaneous Localization and Mapping) [13–15] and so on (Figure 1). There have been an increasing number of applications of pose estimation developed in recent years. Autonomous vehicles use the technology of 6D pose estimation to recognize roads and obstacles. In the factory, the robots use the technology of 6D pose to recognize and grab objects. In the field of augmented reality, 6D pose estimation is used to measure the pose of objects in the real environment and add the virtual objects onto them in a correct pose. Some previous approaches could only detect the object and ensure its position, as is the case for GPS (Global Positioning System) [16] and radar detection [17]. These methods cannot measure the 6D pose of objects accurately. In industrial developments, higher demands are made for new application scenarios. Therefore, 6D pose estimation has become a hot topic in industry in recent years. 6D pose estimation uses a number of kinds of information to solve problems. It obtains texture information, geometric information, and color information to measure the 6D pose of objects. Due to the development of hardware in recent years, depth information is also used frequently in 6D pose estimation. However, 6D pose estimation is faced with many challenges, such as background clutter and inadequate information. Many methods have been proposed to improve the performance and enlarge the range of applications of 6D

**Citation:** He, Z.; Feng, W.; Zhao, X.; Lv, Y. 6D Pose Estimation of Objects: Recent Technologies and Challenges. *Appl. Sci.* **2021**, *11*, 228. https://doi.org/10.3390/app11010228

Received: 11 November 2020 Accepted: 24 December 2020 Published: 29 December 2020

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/ licenses/by/4.0/).

pose estimation, and many new methods have been proposed in recent years. Additionally, there are many challenges in the 6D pose estimation field. A deeper comprehension of these challenges will also help arrive at more practical methods. To understand these methods and challenges more deeply, more detailed classification and performance evaluation need to be carried out. This review summarizes some relevant studies published in recent years and divides these methods into three categories. At the same time, we also analyze the advantages and disadvantages of these categories and challenges in 6D pose estimation.

**Figure 1.** 6D pose estimation applied in bin picking and augmented reality.
