Object distance estimation using a monovision camera

Efemenah Endurance Idohen, Olatunde Mathew Afolayan, Umar Ali Umar


In computer vision, most monovision cameras used for estimating the position of an object only estimate the 2D information of the object without the depth information. Estimating the depth information, which is the distance between the target object and the camera is quite challenging but, in this paper, a less computationally intensive method was used to estimate the object’s distance to complete the 3D information needed to determine the object’s location in cartesian space. In this method, the object was positioned in front of the camera at a sequential distance and was measured directly. The distances measured in the experiment with a set of training data obtained from the image were fitted into a curve using the least-square framework to derive a non-linear function that was used for estimating the object’s distance also known as the z-coordinate. The result from the experiment showed that there was an average error of 1.33 mm between the actual distance and the estimated distance of the object. Hence, this method can be applied in many robotic and autonomous systems applications.


Object Distance estimation; Depth estimation; Distance Measurement; Monovision camera; Pixy2 camera

Full Text:


DOI: http://doi.org/10.11591/ijra.v12i4.pp325-331


Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

IAES International Journal of Robotics and Automation (IJRA)
ISSN 2089-4856, e-ISSN 2722-2586
This journal is published by the Institute of Advanced Engineering and Science (IAES) in collaboration with Intelektual Pustaka Media Utama (IPMU).

Web Analytics Made Easy - Statcounter IJRA Visitor Statistics