Approximation of image boundaries for their classification in the remote sensing of the Earth

1Kosarevych, RYa.
1Karpenko Physico-Mechanical Institute of the National Academy of Science of Ukraine, Lviv, Ukraine
Kosm. nauka tehnol. 1998, 4 ;(4):114–117
https://doi.org/10.15407/knit1998.04.114
Publication Language: Ukrainian
Abstract: 
Boundaries of two-dimensional objects are divided into segments, each being a straight line or an arc. Characteristics of each segment which can be used for the recognition of the object are specified. An algorithm is proposed for a formal description and classification of connected areas. A criterion is suggested for the of recognition objects it is based on the boundary structure analysis.
Keywords: image processing, remote sensing
References: 
1. Kronberg P. Remote study of the Earth, 350 p. (Mir, Moscow, 1988) [in Russian].
2. Pavlidis T. Algorithms for Graphics and Image Processing, 400 p. (Radio i svyaz', Moscow, 1986) [In Russian].
3. Rusyn B. P. Structual Linguistic Real-Time Image Recognition Methods, 128 p. (Nauk. dumka, Kiev, 1986) [In Russian].