Hasil Pencarian  ::  Simpan CSV :: Kembali

"Seiring dengan kemajuan teknologi, kemampuan kendaraan atau robot untuk dapat secara otonom menjelajahi lingkungannya menjadi semakin diminati. Terdapat banyak tantangan yang dihadapi oleh kendaraan atau robot otonom agar dapat melakukannya. Salah satu tantangan tersebut adalah melacak gerakan inkremental dan menganalisis lingkungannya dengan akurat untuk melakukan lokalisasi. Salah satu metode yang dapat digunakan untuk membantu menangani masalah tersebut adalah dengan menggunakan visual odometry. Visual odometry adalah proses mengestimasi gerakan translasi dan rotasi kendaraan atau robot menggunakan kamera yang dipasangkan dengan menganalisa gambar-gambar yang diambil. Dalam penelitian ini, penulis mencoba membangun sebuah sistem visual odometry stereo sederhana. Sistem ini terdiri dari enam bagian utama yaitu mendeteksi fitur dan mengkomputasi deskriptornya menggunakan Oriented FAST and Rotated BRIEF (ORB), mencocokkan fitur secara brute force berdasarkan jarak Hamming dari deskriptor-deskriptor fitur, melacak fitur menggunakan optical flow Lucas-Kanade, melakukan triangulasi terhadap titik-titik fitur menggunakan linear triangulation, mengestimasi translasi dan rotasi dengan menyelesaikan permasalahan Perspective-n-Point (PnP) menggunakan gabungan metode Efficient PnP (EPnP) dan Random Sample Consensus (RANSAC), dan memperbaharui estimasi posisi dan orientasi. Sistem yang dibangun ini memperoleh average translation root mean squared error sebesar 5.1284% dan average rotation error sebesar 0.027 deg/m pada dataset odometry publik KITTI dengan performa kecepatan 18.88 frames per second pada environment komputer 1 core dengan clock speed 2.7 Ghz.

---

As technology advances, the ability of vehicles or robots to be able to autonomously explore their environment is becoming increasingly desirable. There are many challenges that autonomous vehicles or robots face in order to do so. One of the challenges is to track incremental motions and accurately analyze their environment for localization. One of the methods that can be used to help to deal with this problem is by using visual odometry. Visual odometry is the process of estimating the translational and rotational movements of a vehicle or robot using a camera attached by analyzing the images taken. In this research, the author tried to build a simple stereo visual odometry system. This system consists of six main parts, namely detecting features and computing their descriptors using Oriented FAST and Rotated BRIEF (ORB), matching features by brute forcing based on Hamming distance from the feature descriptors, tracking features using Lucas-Kanade optical flow, triangulating the feature points using linear triangulation, estimating translation and rotation by solving Perspective-n-Point (PnP) problems using a combination of Efficient PnP (EPnP) and Random Sample Consensus (RANSAC) methods, and updating the position and orientation estimation. This system has an average translation root mean squared error of 5.1284% and an average rotation error of 0.027 deg/m on the KITTI public odometry dataset with a speed performance of 18.88 frames per second in a 1 core computer with a clock speed of 2.7 Ghz."

"This book describes one of the first systems for sparse point-based 3D reconstruction and egomotion estimation from an image sequence; able to run in real-time at video frame rate and assuming quite weak prior knowledge about camera calibration, motion or scene. Its chapters unify the current perspectives of the robotics and computer vision communities on the 3D vision topic : as usual in robotics sensing, the explicit estimation and propagation of the uncertainty hold a central role in the sequential video processing and is shown to boost the efficiency and performance of the 3D estimation. On the other hand, some of the most relevant topics discussed in SfM by the computer vision scientists are addressed under this probabilistic filtering scheme, namely projective models, spurious rejection, model selection and self-calibration."

"The two volume set LNCS 7431 and 7432 constitutes the refereed proceedings of the 8th International Symposium on Visual Computing, ISVC 2012, held in Rethymnon, Crete, Greece, in July 2012. The 68 revised full papers and 35 poster papers presented together with 45 special track papers were carefully reviewed and selected from more than 200 submissions. The papers are organized in topical sections, Part I (LNCS 7431) comprises computational bioimaging, computer graphics, calibration and 3D vision, object recognition, illumination, modeling, and segmentation, visualization, 3D mapping, modeling and surface reconstruction, motion and tracking, optimization for vision, graphics, and medical imaging, HCI and recognition. Part II (LNCS 7432) comprises topics such as unconstrained biometrics, advances and trends, intelligent environments, algorithms and applications; applications, virtual reality, face processing and recognition."