Miguel Realpe - Academia.edu (original) (raw)

Papers by Miguel Realpe

Abstract. Recent stereo cameras provide reliable 3D reconstructions. These are useful for selecti... more Abstract. Recent stereo cameras provide reliable 3D reconstructions. These are useful for selecting ground-plane points, register them and building mosaics of cluttered ground planes. In this paper we propose a 2D Iterated Closest Point (ICP) registration method, based on the distance transform, combined with a fine-tuning-registration step using directly the image data. Experiments with real data show that ICP is robust to 3D reconstruction differences due to motion and the fine tuning step minimizes the effect of the uncertainty in the 3D reconstructions. 1

El presente trabajo pretende solucionar los problemas de localizacion y mapeo del ambiente de nav... more El presente trabajo pretende solucionar los problemas de localizacion y mapeo del ambiente de navegacion de unrobot a partir de imagenes 3D a color .Se ha realizado la implementacion de un sistema de registro de imagenestridimensionales mediante el algoritmo ICP, el cual toma ventaja de la informacion de color de las imagenes 3D ydetermina la localizacion del robot en un momento determinado. Ademas se ha realizado la alineacion de todas lasimagenes adquiridas por el robot, con lo cual se produce un mapa tridimensional del ambiente por el cual navega.Resultados experimentales usando tres tipos de recorridos con un robot fueron realizados.

... Hidden Markov models (HMMs) consiste en una máquina de estados estocásticos que permite un an... more ... Hidden Markov models (HMMs) consiste en una máquina de estados estocásticos que permite un análisis sofisticado de datos variables en el ... En esta propuesta se emplean técnicas de detección y seguimiento de puntos característicos de personas que dan como resultado ...

This paper describes the implementation of an open-source software developed using Python, which ... more This paper describes the implementation of an open-source software developed using Python, which facilitates the materials selection process commonly used in engineering. This software has been developed by non-CS students (Materials Engineering, Food Engineering and Chemistry Engineering), as a project course of their 1st-year cross-curricular course of CS1 (“Programming Fundamentals”), in order to connect their CS1 learning process with core subjects related to their careers, aiming to motivate both, the use of computer programming in their personal development and also, their interest in their professional career. The program developed allows choosing between different types of materials, based on specific characteristics required by the user; furthermore, this program enables the visualization of the Michael Ashby methodology for materials selection, which allows non-CS students to solve a problem related to their career, while it gives upper-level students a new tool to learn i...

El presente trabajo ha sido desarrollado como un proyecto de tesis de graduacion de la FIEC [1]. ... more El presente trabajo ha sido desarrollado como un proyecto de tesis de graduacion de la FIEC [1]. El principal objetivo es generar mapas tridimensionales de ambientes reales a partir de imagenes de profundidad que son adquiridas desde una camara de estereo vision montada sobre un robot movil que explora dichos ambientes y, a la vez, estimar la localizacion del robot dentro del ambiente. Debido a la complejidad y para facilitar su implementacion, se ha decidido dividir en proyecto en tres fases que son: sistema de adquisicion 3D, sistema de localizacion, sistema de mapeo 3D.

In recent years 3D urban maps have become more common, thus providing complex point clouds that i... more In recent years 3D urban maps have become more common, thus providing complex point clouds that include diverse urban furniture such as pole-like objects. Utility poles detection in urban environment is of particular interest for electric utility companies in order to maintain an updated inventory for better planning and management. The present study develops an automatic method for the detection of utility poles from noisy point cloud data of Guayaquil - Ecuador, where many poles are located very close to buildings, which increases the difficulty of discriminating poles, walls, columns, fences and building corners. The proposed method applies a segmentation stage based on clustering with vertical voxels and a classification stage based on neural networks.

In this paper we discuss the feasibility of using a DCNN to implement Fall and Head detection for... more In this paper we discuss the feasibility of using a DCNN to implement Fall and Head detection for a Danger Management System. For this propose, AlexNet and Inception-v3 models in MatConvNet and TensorFlow libraries were used for training new DCNNs with Fine-Tuning method. Additionally, a new public dataset was created, which includes diverse fall poses, as well as, top views of people walking in a scene.

The problem of loss of line of sight when operating drones has become a reality with adverse effe... more The problem of loss of line of sight when operating drones has become a reality with adverse effects for professional and amateur drone operators, since it brings technical problems such as loss of data collected by the device in one or more instants of time during the flight and even misunderstandings of legal nature when the drone flies over prohibited or private places. This paper describes the implementation of a drone monitoring system using the Internet as a long-range communication network in order to avoid the problem of loss of communication between the ground station and the device. For this, a simulated environment is used through an appropriate open software tool. The operation of the system is based on a client that makes requests to a server, the latter in turn communicates with several servers, each of which has a drone connected to it. In the proposed system when a drone is ready to start a flight, its server informs the main server of the system, which in turn gives feedback to the client informing it that the device is ready to carry out the flight; this way customers can send a mission to the device and keep track of its progress in real time on the screen of their web application.

It is important for farmers to know the level of chlorophyll in plants since this depends on the ... more It is important for farmers to know the level of chlorophyll in plants since this depends on the treatment they should give to their crops. There are two common classic methods to get chlorophyll values: from laboratory analysis and electronic devices. Both methods obtain the chlorophyll level of one sample at a time, although they can be destructive. The objective of this research is to develop a system that allows for obtaining the chlorophyll level of plants using multispectral images.

Proceedings of the 17th LACCEI International Multi-Conference for Engineering, Education, and Technology: “Industry, Innovation, and Infrastructure for Sustainable Cities and Communities”, 2019

Drones or unmanned aerial vehicle (UAV) are very useful for image acquisition. However, all the i... more Drones or unmanned aerial vehicle (UAV) are very useful for image acquisition. However, all the images acquired during a flight must be combined in order to create valuable information about a terrain or crop. There are many programs that combine images (image registration), each one with different properties (performance, precision, price, etc.). In this project some Open Source programs were analyzed in order to define the best software for Ecuadorian small and medium-sized farmers. The results obtained could be of interest for software developers since they may modify and integrate the programs into a simpler workflow. Additionally, it allows to reduce costs since no license payment are needed, which may motive small and medium-sized farmers. A public repository has been created as part of the results of this project, which includes code for pre-processing multispectral images and code for creating RGB, CIR and NDVI fused maps.

Examen de FUNDAMENTOS DE ROBÓTICA del 2010-2S de la 1° Evaluación

Proceedings of the 19th LACCEI International Multi-Conference for Engineering, Education, and Technology: “Prospective and trends in technology and skills for sustainable social development” “Leveraging emerging technologies to construct the future”

Journal of Automation and Control Engineering

Driverless vehicles are currently being tested on public roads in order to examine their ability ... more Driverless vehicles are currently being tested on public roads in order to examine their ability to perform in a safe and reliable way in real world situations. However, the long-term reliable operation of a vehicle's diverse sensors and the effects of potential sensor faults in the vehicle system have not been tested yet. This paper is proposing a sensor fusion architecture that minimizes the influence of a sensor fault. Experimental results are presented simulating faults by introducing displacements in the sensor information from the KITTI dataset. 

2016 35th Chinese Control Conference (CCC), 2016

ABSTRACT El presente trabajo ha sido desarrollado como un proyecto de tesis de graduación de la F... more ABSTRACT El presente trabajo ha sido desarrollado como un proyecto de tesis de graduación de la FIEC [1]. El principal objetivo es generar mapas tridimensionales de ambientes reales a partir de imágenes de profundidad que son adquiridas desde una cámara de estéreo visión montada sobre un robot móvil que explora dichos ambientes y, a la vez, estimar la localización del robot dentro del ambiente. Debido a la complejidad y para facilitar su implementación, se ha decidido dividir en proyecto en tres fases que son: sistema de adquisición 3D, sistema de localización, sistema de mapeo 3D.

2015 IEEE 7th International Conference on Cybernetics and Intelligent Systems (CIS) and IEEE Conference on Robotics, Automation and Mechatronics (RAM), 2015

Lecture Notes in Computer Science, 2007

Recent stereo cameras provide reliable 3D reconstructions. These are useful for selecting ground-... more Recent stereo cameras provide reliable 3D reconstructions. These are useful for selecting ground-plane points, register them and building mosaics of cluttered ground planes. In this paper we propose a 2D Iterated Closest Point (ICP) registration method, based on the distance transform, combined with a fine-tuning-registration step using directly the image data. Experiments with real data show that ICP is robust to 3D reconstruction differences due to motion and the fine tuning step minimizes the effect of the uncertainty in the 3D reconstructions.