深度感測器與慣性測量單元之座標校正方法

Method of Calibrating between Depth Sensor and Inertial Measurement Unit

黃繼德(Ji-De Huang)

《電腦與通訊》 153期 (2013/10) Pp. 138-145

摘要

深度感測器（如Kinect與雷射掃描儀），常常搭配慣性感測器（如加速計與陀螺儀）應用於機器人領域。這兩種感測器皆能偵測3D運動狀態，但各有各自的座標定義，因而於實際應用時，需仔細的進行角度校正來達到座標對準；若無適當校正，將會從兩種感測器計算出不一致的運動狀態，造成後續處理失敗。翻閱文獻，雖然現在已有影像感測器與慣性感測器的校正演算法，但卻無深度感測器與慣性感測器之校正演算法。本篇論文便在說明如何分析這兩種感測器的資料，進而計算出校正矩陣；此外亦介紹自行開發之高準確度慣性感測模組。

關鍵字

深度感測器 ； 慣性感測器 ； 微控制器

並列摘要

Depth Sensor (ex: Kinect and Laser Scanner) and IMU (ex: accelerometer and gyroscope) are often combined together in robotic application. Both types of sensor can detect 3D motion, and each has its definition of coordinate. Without appropriate aligning, one will get different motion from each sensor, and make following processing fail. For this reason, one should carefully align both sensors' coordinate. Although there are papers dealing with calibrating between image sensor and IMU, there still no method for calibrating between depth sensor and IMU. This paper details the method to analyze data from both sensor, and calculate the calibration matrix. Besides, self-made high precision IMU is also descripted in this paper.

並列關鍵字

Depth Sensor ； Inertial Measurement Unit ； IMU ； Microcontroller ； MCU