Open AccessPrediction of XYZ coordinates from an image using mono camera
Author(s) -
Muslikhin Muslikhin,
Dessy Irmawati,
Fatchul Arifin,
Aris Nasuha,
Nur Hasanah,
Yuniar Indrihapsari
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1456/1/012015
Subject(s) - position (finance) , artificial intelligence , computer vision , computer science , centroid , object (grammar) , image (mathematics) , point (geometry) , stereo camera , homogeneous , class (philosophy) , simple (philosophy) , mathematics , geometry , finance , combinatorics , economics , philosophy , epistemology
Estimating the position of a homogeneous object from an image for XY position is quite simple because it has the same dimensions XY. However, determining the XYZ position requires a unique approach. Generally, for estimating 3D position, stereo camera or expensive cameras are used with complicated computer vision algorithms. In this paper, we classify the position of an object using a mono camera. The image is divided into 3185 classes and five layers as a machine learning algorithm references. The k-nearest neighbors (kNN) approach usually is to find the closest point of the centroids to the closest class. Thus, this approach can be used as a three-axis prediction method that can afford the best performance solution.