Mensuração da velocidade de deslocamento de corpos rígidos em tempo real por análise de sequência de imagens

• Ano de defesa: 2013
• Autor(a) principal: Filitto, Danilo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual de Maringá; Brasil; Departamento de Informática; Programa de Pós-Graduação em Ciência da Computação; UEM; Maringá, PR; Centro de Tecnologia
• Palavras-chave em Português: Segmentação de imagem; Detecção de movimento; Modelos de transformação geométrica; Mensuração da velocidade; Corpos rígidos; Brasil.; Segmentation; Detection; Geometric transformation models; Measurement speed; Rigid bodies; Brazil.; Ciências Exatas e da Terra; Ciência da Computação
• Link de acesso: http://repositorio.uem.br:8080/jspui/handle/1/2534

This work demonstrates a methodology aimed to measure the velocity of a rigid body in motion, taking as a starting point, a sequence of images taken by a digital camera. This process requires the measurement of the speed of execution fundamental steps: 1) detect and segment the moving bodies in the scene image and, 2) turning points of the body known belonging to the desired image plane (2D) object into space (3D). For step detection and segmentation, this work implements two techniques - one based on time average values of each pixel recorded in N image frames, and another based on historical values associated with N pixels recorded in picture frames . After cleavage of the bodies, the transformation is performed to obtain the correspondence points belonging to the images of bodies with desired points in space corresponding real object, such a transformation may be performed using the collinearity equations, the model linear transformation directly, by polynomial transformation model, the model of affine transformation or transformation model of rational functions. Obtaining those points in the object space allows to calculate the actual distance the body traveled in a time interval automatically timed. These two values (space and time) to calculate the speed with which the body moves. To validate the proposed methodology was performed a rigorously controlled experiment that involved the use of a device called air rail, five photoelectric sensors scattered about the same with known distances between themselves and a timer connected to the sensors. At the time when a moving body passes a sensor, the timer is stopped and thus, has the time to take in the body travels a distance from the origin to each of the sensors disposed on the tracks of air. This entire experiment was filmed and on the frames of this film, we applied the proposed methodology to calculate the speed of a body only by the use of image sequences. The results obtained with the use of sensors are then compared with the results calculated by use of the images.