Skip to Main content Skip to Navigation
Journal articles

Unsupervised marked point process model for boat extraction and counting in harbors from high resolution optical remotely sensed images

Abstract : Marked point process models have been successfully applied to object extraction in high resolution optical remotely sensed images during the last ten years. The models typically consist of two types of energy terms : a data term which reflects the fidelity of the configuration to the input image and a prior term which incorporates some knowledge about the objects to be extracted. In this paper we deal with the problem of extracting boats in harbors. This is a difficult problem due to the particular distribution of the objects in this case. We describe a previously developed marked point process model of ellipses for such a goal, for which we automatically determine one of its key parameters (the direction of the boats). We present the drawbacks of the model due to the hard constraints imposed, which we then relax and propose a new, more general model.
Document type :
Journal articles
Complete list of metadata

Cited literature [23 references]  Display  Hide  Download

https://hal.inria.fr/hal-01119448
Contributor : Paula Craciun <>
Submitted on : Monday, February 23, 2015 - 11:35:08 AM
Last modification on : Saturday, January 27, 2018 - 1:31:40 AM
Long-term archiving on: : Wednesday, May 27, 2015 - 9:46:19 AM

File

RFTP_eng.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01119448, version 1

Collections

Citation

Paula Craciun, Josiane Zerubia. Unsupervised marked point process model for boat extraction and counting in harbors from high resolution optical remotely sensed images. Revue Française de Photogrammétrie et de Télédétection, Société Française de Photogrammétrie et de Télédétection, 2014, 207, pp.33-44. ⟨hal-01119448⟩

Share