Interciência & Sociedade53 NAVEGAÇÃO AUTÔNOMA EM AMBIENTES AGRÍCOLAS UTILIZANDO VISÃO ESTÉREO RESUMO: - - PALAVRAS-CHAVE: - turais. ABSTRACT: possible to check if a predetermined path can be used to reach the designed goal and estimate for how long this path stand as valid, without further new calculations. With the purpose of collaborate with the correspondence method results, responsible for build the disparity matrix, the captured images in each camera are recti ed, so both of them can be aligned hori ontally. ased on this matrix, the v disparity matrix is created, so nally it is possible to map the depth of the ob ects on the scene, which can repre- sent obstructions along the path. KEYWORDS: tereo vision, stereo correspondence, recti cation, v-disparity, natural scenes. 1. INTRODUÇÃO - - - - uma autonomia bastante grande conside- Interciência & Sociedade54 Considerando-se então um am- - - - - - - - caminho. - - - - - - - - balho descrito neste artigo. 2. METODOLOGIA DE DESENVOLVIMEN- TO - - - - - cena, gerando assim duas imagens distin- tas dessa mesma cena, sendo uma imagem - - - - - - Figura 1 - Interciência & Sociedade55 - - - encontrados em um sistema monocular, O ambiente selecionado como es- - - Figura 2 - - pixels . cena. - - - - - - - - - - - - - tes. - - - - Interciência & Sociedade56 - - guinte forma: - - balho, são utilizadas as intersec- - - - as coordenadas encontradas no então as matrizes de rotação e - - - dência. - - - - - Figura 3 - - - - Interciência & Sociedade57 tentam relacionar um bloco de uma deter- bloco da imagem resulta na matriz de dis- De uma forma geral, os blocos são - - - - - - - ambiente. - pixel a pixel, o algoritmo de cor- - - - - - - - - - Figura 4 - foi utilizado um bloco de tamanho igual a pixels - - Em um caso ideal, dois blocos - - - Figura 5 - Interciência & Sociedade58 - - - - - - - em azul. Uma indicação da direção a ser - Figura 6 - 3. RESULTADOS - - - - cessamento. Figura 7 - Interciência & Sociedade59 - - Figura 8 - - um obstáculo na região central da imagem. - - - - - Figura 9 - Interciência & Sociedade60 - - - - - - - - - de acordo com os obstáculos encontra- a Figura 8 e a Figura 9, sugerem a re- - não será uma boa decisão mais a frente. - encontrado a frente. Figura 10 - Na Figura 11 ocorre a situação - - - - nhum obstáculo foi detectado. Interciência & Sociedade61 4. CONCLUSÃO As imagens obtidas em um am- - minação e efeito de condensação das lentes - - - - - Figura 11 - - - - - - 5. REFERÊNCIAS BIBLIOGRÁFICAS Block Matching Algorithms For Mo- tion Estimation Learning OpenCV: Computer Vision with OpenCV Library Interciência & Sociedade62 Path and Obstacle Detection Using Decision Ne- IEEE Transactions on In- telligent Transportation Systems - Blo- ck Matching for Object Tracking of California. Space Research Institute Russia. Proceedings International Symposium on Experi- mental Robotics (ISER) - tation. Intelligent Vehicle Symposium, 2002. IEEE - mo em Ambientes com Obstáculos. XVIII Congresso Brasileiro de Automática Navegação de ro- bôs móveis utilizando visão estéreo - sidade de São Paulo. - 2007 IEEE In- telligent Vehicles Symposium Correspondência densa para sistemas de visão estereoscópica para robótica móvel - rina. VI Workshop de Visão Computacional - sidente Prudente. Interciência & Sociedade63 Vinicius Bergoli Trentini: Victor de Assis Rodrigues: Aparecido Nilceu Marana: Luciano Cássio Lúlio: USP. Mário Luiz Tronco: