Depth perception processing using a monocular image

• Main Authors: Carandang, Leandro Angelico T., Doroja, Marc R., Sandoval, Federico N., Zialcita, Gabriel Niccolo C.
• Format: text
• Language: English
• Published: Animo Repository 2012
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/12173
• Institution: De La Salle University

Images are the most commonly utilized representation of our surroundings these images are in 2D. However, modern day computers currently have no ability to identify the relative depth of the objects within the image, thus not being able to identify the objects in the image. Depth perception is a common problem in research because there are many ways to tackle the problem such as using depth maps, using stereoscopic devices, etc. Depth perception is important, because depth perception involves the ability to interact with the environment. Today 2D image processing is prominently used in Android and iOS as measurement tools that can measure the height of an object in a 2D image and calories present in a certain food in a 2D image. Our group used existing algorithms to design a method that takes advantage of monocular depth cues present in a 2D image to get an understanding of the relative depth of objects within the image using techniques such as edge detection, line detection, binarization, gray scaling, blob detection, object detection, etc. This allowed the computer to identify the layering of the depth in an image, such as if object A is in front of the Object B, or Object B is in front of Object A, or if another entire object Object C is either behind or in front of both objects. This enabled a better way of interacting with the environment using machines.