Mais informações

TATEBE, Kenji; TSUJIMOTO, Makoto. Application of digital image processing techniques in the analysis of pedestrian behavior. In: NUTAU: TECNOLOGIA E DESENVOLVIMENTO, 3., 1997, São Paulo. Anais… São Paulo: USP, 1997. p. 241-254.
Clique no nome do(s) autor(es) para ver o currículo Lattes:

Dados do autor na base InfoHab:
Número de Trabalhos: 3 (Nenhum com arquivo PDF disponível)
Citações: 3
Índice h: 1  
Co-autores: 2

Dados do autor na base InfoHab:
Número de Trabalhos: 1 (Nenhum com arquivo PDF disponível)
Citações: Nenhuma citação encontrada
Índice h: Indice h não calculado  
Co-autores: Nenhum co-autor encontrado

Abstract

In trying to improve city and building conditions for pedestrians, it is necessary to comprehend how the people use them. One effective means is videotape recordings of the human behavior in cities. However, until now, the quantitative analysis of visual data has involved some problems, a lot of working hours, the subjective judgement of the experimenter, and so on. So methods by which the visual data can be objectively and automatically analyzed should be developed. A method which analyzes pedestrian movement quantitatively by using digital image processing techniques is discussed in this paper. The results are summarized as follows: 1)The method was developed to measure the position of pedestrians automatically 30 times in every second. The main system of this method, using the fact that the hair is normally darker than the other body parts, is to trace the position of head hair automatically in the binary image which is made from a videotape recording of pedestrian movement. 2)Two experiments on people walking a) in a straight path and, b) round an L-shaped corner, were performed. By using the new method the data of both experiments for position, velocity and acceleration of pedestrians was processed. It was found to be effective in analyzing the body movement of pedestrians during the cyclic acceleration changes in their natural walking rhythm.
-