Abstract
A computer vision system was developed that allows calculating the correspondence relationship
between a robotic assembly table and an image captured by a tilted webcam, obtaining a
pixel/centimeter ratio. The system initially captures a color image in RGB format in the visible
spectrum with a resolution of 640x480 pixels, then applies a histogram equalization and
performs a transformation from the RGB plane to the HSV plane, followed by applying red color
filtering to a Rectangular area with thickness at its edges of approximately 1 cm outwards with
rectangular dimensions of 33.2 cm by 23.2 cm which we will call the work area and is
superimposed on a 90 cm by 50 cm metal table.
Once the image is binarized, an edge detection is applied and later a four-sided polygonal
approximation to identify the four corners of the work area, this process results in two rectangles,
one internal and one external, the one with the smallest is taken. area. After this, a perspective
transformation is performed to obtain an image equivalent to having the camera in a top view
position. Experimentally, errors were found in the upper areas of the work area, for this reason
this research proposes a post-processing that serves as a final correction to calculate the proposed
pixel/centimeter ratio
between a robotic assembly table and an image captured by a tilted webcam, obtaining a
pixel/centimeter ratio. The system initially captures a color image in RGB format in the visible
spectrum with a resolution of 640x480 pixels, then applies a histogram equalization and
performs a transformation from the RGB plane to the HSV plane, followed by applying red color
filtering to a Rectangular area with thickness at its edges of approximately 1 cm outwards with
rectangular dimensions of 33.2 cm by 23.2 cm which we will call the work area and is
superimposed on a 90 cm by 50 cm metal table.
Once the image is binarized, an edge detection is applied and later a four-sided polygonal
approximation to identify the four corners of the work area, this process results in two rectangles,
one internal and one external, the one with the smallest is taken. area. After this, a perspective
transformation is performed to obtain an image equivalent to having the camera in a top view
position. Experimentally, errors were found in the upper areas of the work area, for this reason
this research proposes a post-processing that serves as a final correction to calculate the proposed
pixel/centimeter ratio
| Translated title of the contribution | Correspondence System for Pixel-Centimeter Ratio Calculation in an Assembly Table Using a Tilted Webcam |
|---|---|
| Original language | Spanish |
| Title of host publication | CIENCIA, TECNOLOGÍA E INNOVACIÓN: TECNOLOGÍAS APLICADAS |
| Editors | Maria Alexandra Fajardo Perdomo , Aldemas Junior Mendoza Ramírez, Juan David Jiménez Miranda |
| Place of Publication | Bogotá, Colombia |
| Publisher | Editorial Universidad INCCA |
| Chapter | 2 |
| Pages | 41-61 |
| Number of pages | 21 |
| Volume | 1 |
| Edition | 1 |
| ISBN (Print) | 978-958-9009-83-3 |
| State | Published - 30 Aug 2024 |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 9 Industry, Innovation, and Infrastructure
Fingerprint
Dive into the research topics of 'Correspondence System for Pixel-Centimeter Ratio Calculation in an Assembly Table Using a Tilted Webcam'. Together they form a unique fingerprint.Projects
- 1 Finished
-
Aproximación a una arquitectura cognitiva para el aprendizaje y generalización de un proceso de clasificación aplicado en un robot colaborativo
Martinez-Alvarez, A. (PI), Contreras Pérez, J. D. (CoI), Loaiza Correa, H. (CoI), Mosquera De La Cruz, J. H. (CoI) & Nope Rodríguez, S. E. (CoI)
01/02/22 → 30/06/23
Project: Research
Research output
- 1 Article
-
Sistema de interacción humano-robot para la enseñanza-aprendizaje de una tarea de ordenamiento de objetos mediante comunicación verbal y gestual
Martinez, A., Nope Rodríguez, S. E., Loaiza, H., Mosquera-Delacruz, J. H., Rodriguez, G. & Jamioy-Cabrera, J. D., 29 Nov 2023, In: Ingeniería y Competitividad. 25, Supl, p. 1-21 21 p., e-20613133.Translated title of the contribution :Human-robot interaction system for teaching-learning of an object sorting task by means of verbal and gestural communication Research output: Contribution to journal › Article › peer-review
Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver