RESUMO
The acquisition of the body temperature of animals kept in captivity in biology laboratories is crucial for several studies in the field of animal biology. Traditionally, the acquisition process was carried out manually, which does not guarantee much accuracy or consistency in the acquired data and was painful for the animal. The process was then switched to a semi-manual process using a thermal camera, but it still involved manually clicking on each part of the animal's body every 20 s of the video to obtain temperature values, making it a time-consuming, non-automatic, and difficult process. This project aims to automate this acquisition process through the automatic recognition of parts of a lizard's body, reading the temperature in these parts based on a video taken with two cameras simultaneously: an RGB camera and a thermal camera. The first camera detects the location of the lizard's various body parts using artificial intelligence techniques, and the second camera allows reading of the respective temperature of each part. Due to the lack of lizard datasets, either in the biology laboratory or online, a dataset had to be created from scratch, containing the identification of the lizard and six of its body parts. YOLOv5 was used to detect the lizard and its body parts in RGB images, achieving a precision of 90.00% and a recall of 98.80%. After initial calibration, the RGB and thermal camera images are properly localised, making it possible to know the lizard's position, even when the lizard is at the same temperature as its surrounding environment, through a coordinate conversion from the RGB image to the thermal image. The thermal image has a colour temperature scale with the respective maximum and minimum temperature values, which is used to read each pixel of the thermal image, thus allowing the correct temperature to be read in each part of the lizard.
Assuntos
Inteligência Artificial , Temperatura Corporal , Lagartos , Animais , Lagartos/fisiologia , Temperatura Corporal/fisiologia , Gravação em Vídeo/métodos , Processamento de Imagem Assistida por Computador/métodosRESUMO
In this work, a large-scale tactile detection system is proposed, whose development is based on a soft structure using Machine Learning and Computer Vision algorithms to map the surface of a forearm sleeve. The current application has a cylindrical design, whose dimensions intend to be like a human forearm or bicep. The model was developed assuming that deformations occur only at one section at a time. The goal for this system is to be coupled with the CHARMIE robot, a collaborative robot for domestic and medical environments. This system allows the contact detection of the entire forearm surface enabling interaction between a Human Being and a robot. A matrix with sections can be configured to present certain functionalities, allowing CHARMIE to detect contact in a particular section, and thus perform a specific behaviour. After building the dataset, an Artificial Neural Network (ANN) was created. This network was called Section Detection Network (SDN), and through Supervised Learning, a model was created to predict the contact location. Furthermore, Stratified K-Fold Cross Validation (SKFCV) was used to divide the dataset. All these steps resulted in Neural Network with a test data accuracy higher than 80%. Regarding the real-time evaluation, a graphical interface was structured to demonstrate the predicted class and the corresponding probability. This research concluded that the method described has enormous potential to be used as a tool for service robots allowing enhanced human-robot interaction.
Assuntos
Robótica , Humanos , Robótica/métodos , Tato , Redes Neurais de Computação , Algoritmos , Aprendizado de Máquina SupervisionadoRESUMO
This paper presents a road surface scanning system that operates with a trichromatic line scan camera with light emitting diode (LED) lighting achieving road surface resolution under a millimeter. It was part of a project named Roadkills-Intelligent systems for surveying mortality of amphibians in Portuguese roads, sponsored by the Portuguese Science and Technology Foundation. A trailer was developed in order to accommodate the complete system with standalone power generation, computer image capture and recording, controlled lighting to operate day or night without disturbance, incremental encoder with 5000 pulses per revolution attached to one of the trailer wheels, under a meter Global Positioning System (GPS) localization, easy to utilize with any vehicle with a trailer towing system and focused on a complete low cost solution. The paper describes the system architecture of the developed prototype, its calibration procedure, the performed experimentation and some obtained results, along with a discussion and comparison with existing systems. Sustained operating trailer speeds of up to 30 km/h are achievable without loss of quality at 4096 pixels' image width (1 m width of road surface) with 250 µm/pixel resolution. Higher scanning speeds can be achieved by lowering the image resolution (120 km/h with 1 mm/pixel). Computer vision algorithms are under development to operate on the captured images in order to automatically detect road-kills of amphibians.