当前位置： SCI文献检索 > ENVIRONMENTAL MANAGEMENT期刊下所有文献 > Object-based classification as an alternative approach to the traditional pixel-based classification to identify potential habitat of the grasshopper sparrow.

Object-based classification as an alternative approach to the traditional pixel-based classification to identify potential habitat of the grasshopper sparrow.

Abstract：

:The traditional method of identifying wildlife habitat distribution over large regions consists of pixel-based classification of satellite images into a suite of habitat classes used to select suitable habitat patches. Object-based classification is a new method that can achieve the same objective based on the segmentation of spectral bands of the image creating homogeneous polygons with regard to spatial or spectral characteristics. The segmentation algorithm does not solely rely on the single pixel value, but also on shape, texture, and pixel spatial continuity. The object-based classification is a knowledge base process where an interpretation key is developed using ground control points and objects are assigned to specific classes according to threshold values of determined spectral and/or spatial attributes. We developed a model using the eCognition software to identify suitable habitats for the Grasshopper Sparrow, a rare and declining species found in southwestern Québec. The model was developed in a region with known breeding sites and applied on other images covering adjacent regions where potential breeding habitats may be present. We were successful in locating potential habitats in areas where dairy farming prevailed but failed in an adjacent region covered by a distinct Landsat scene and dominated by annual crops. We discuss the added value of this method, such as the possibility to use the contextual information associated to objects and the ability to eliminate unsuitable areas in the segmentation and land cover classification processes, as well as technical and logistical constraints. A series of recommendations on the use of this method and on conservation issues of Grasshopper Sparrow habitat is also provided.

journal_name

Environ Manage

journal_title

Environmental management

authors

Jobin B,Labrecque S,Grenier M,Falardeau G

doi

10.1007/s00267-007-9031-0

subject

Has Abstract

pub_date

2008-01-01 00:00:00

pages

20-31

issue

eissn

0364-152X

issn

1432-1009

journal_volume

pub_type

杂志文章

在线工具

Object-based classification as an alternative approach to the traditional pixel-based classification to identify potential habitat of the grasshopper sparrow.