Scale parameter selection by spatial statistics for GeOBIA: Using mean-shift based multi-scale segmentation as an example

Ming, Dongping; Li, Jonathan; Wang, Junyi; Zhang, Min

doi:10.1016/j.isprsjprs.2015.04.010

Cited by 141 publications

(80 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This is an inherent problem in any expert system, although in the case of GEOBIA, research into statistical methods such as the Estimation of Scale Parameter [47] that minimize within object local variance while maximizing between object variance provide some guidance in this regard. Other parameter optimization techniques [48][49][50] have been recently introduced and the ESP technique was automated and updated to include multidimensional data [51], and these would likely improve the segmentation process and may obviate the need for using the spectral difference merging of size-limited image object that otherwise have similar characteristic features to neighboring objects.…”

Section: Discussionmentioning

confidence: 99%

An Object-Oriented Approach to Extracting Productive Fossil Localities from Remotely Sensed Imagery

et al. 2015

View full text Add to dashboard Cite

† These authors contributed equally to this work.Abstract: Most vertebrate fossils are rare and difficult to find and although paleontologists and paleoanthropologists use geological maps to identify potential fossil-bearing deposits, the process of locating fossiliferous localities often involves a great deal of luck. One way to reduce the role of serendipity is to develop predictive models that increase the likelihood of locating fossils by identifying combinations of geological, geospatial, and spectral features that are common to productive localities. We applied GEographic Object-Based Image Analysis (GEOBIA) of high resolution QuickBird and medium resolution images from the Landsat 8 Operational Land Imager (OLI) along with GIS data such as slope and surface geology layers to identify potentially productive Eocene vertebrate fossil localities in the Great Divide Basin, Wyoming. The spectral and spatial characteristics of the image objects that represent a highly productive locality (WMU-VP-222) were used to extract similar image objects in the area covered by the high resolution imagery and throughout the basin using the Landsat imagery. During the 2013 summer field season, twenty-six locations that would not have been spotted from the road in a traditional ground survey were visited. Fourteen of the eighteen localities that were fossiliferous were identified by the predictive model. In 2014, the GEOBIA techniques were applied to Landsat 8 imagery of the entire basin, correctly identifying six new productive localities in a previously unsurveyed part of the basin.

show abstract

Section: Discussionmentioning

confidence: 99%

An Object-Oriented Approach to Extracting Productive Fossil Localities from Remotely Sensed Imagery

et al. 2015

View full text Add to dashboard Cite

show abstract

“…A large number of pixel features and object features have been developed, and a series of scale parameter evaluation tools specific to the object method has been promoted; for example, a tool to estimate scale parameter for multiresolution image segmentation of remotely sensed data (ESP) [44], spatial statistics [45], and semi-automatic optimization scheme have been used for object-oriented classification [46]. Classification accuracy can be doubled by introducing these features and tools into the classification process for different complex scenes [4,[47][48][49][50][51].…”

Section: Remote Sensing Data Sources Referencesmentioning

confidence: 99%

A Review of Fine-Scale Land Use and Land Cover Classification in Open-Pit Mining Areas by Remote Sensing Techniques

Chen

et al. 2017

Remote Sensing

View full text Add to dashboard Cite

Abstract:Over recent decades, fine-scale land use and land cover classification in open-pit mine areas (LCCMA) has become very important for understanding the influence of mining activities on the regional geo-environment, and for environmental impact assessment procedure. This research reviews advances in fine-scale LCCMA from the following aspects. Firstly, it analyzes and proposes classification thematic resolution for LCCMA. Secondly, remote sensing data sources, features, feature selection methods, and classification algorithms for LCCMA are summarized. Thirdly, three major factors that affect LCCMA are discussed: significant three-dimensional terrain features, strong LCCMA feature variability, and homogeneity of spectral-spatial features. Correspondingly, three key scientific issues that limit the accuracy of LCCMA are presented. Finally, several future research directions are discussed: (1) unitization of new sensors, particularly those with stereo survey ability; (2) procurement of sensitive features by new sensors and combinations of sensitive features using novel feature selection methods; (3) development of robust and self-adjusted classification algorithms, such as ensemble learning and deep learning for LCCMA; and (4) application of fine-scale mining information for regularity and management of mines.

show abstract

“…Multi-scale image segmentation is the foundational procedure of OBIA in which the digital image is transformed from discrete pixels into spectrally homogeneous, contiguous image object primitives [24]. Various segmentation techniques have been developed with different results, and the selection of the image segmentation technique may lead to different classification accuracies in OBIA [25].…”

Section: Multi-resolution Segmentationmentioning

confidence: 99%

Land Cover Change Detection in Urban Lake Areas Using Multi-Temporary Very High Spatial Resolution Aerial Images

Zhang

Tan

Zheng

et al. 2018

Water

View full text Add to dashboard Cite

Abstract:The availability of very high spatial resolution (VHR) remote sensing imagery provides unique opportunities to exploit meaningful change information in detail with object-oriented image analysis. This study investigated land cover (LC) changes in Shahu Lake of Wuhan using multi-temporal VHR aerial images in the years 1978, 1981, 1989, 1995, 2003, and 2011. A multi-resolution segmentation algorithm and CART (classification and regression trees) classifier were employed to perform highly accurate LC classification of the individual images, while a post-classification comparison method was used to detect changes. The experiments demonstrated that significant changes in LC occurred along with the rapid urbanization during 1978-2011. The dominant changes that took place in the study area were lake and vegetation shrinking, replaced by high density buildings and roads. The total area of Shahu Lake decreased from~7.64 km 2 to~3.60 km 2 during the past 33 years, where 52.91% of its original area was lost. The presented results also indicated that urban expansion and inadequate legislative protection are the main factors in Shahu Lake's shrinking. The object-oriented change detection schema presented in this manuscript enables us to better understand the specific spatial changes of Shahu Lake, which can be used to make reasonable decisions for lake protection and urban development.

show abstract

Scale parameter selection by spatial statistics for GeOBIA: Using mean-shift based multi-scale segmentation as an example

Cited by 141 publications

References 44 publications

An Object-Oriented Approach to Extracting Productive Fossil Localities from Remotely Sensed Imagery

An Object-Oriented Approach to Extracting Productive Fossil Localities from Remotely Sensed Imagery

A Review of Fine-Scale Land Use and Land Cover Classification in Open-Pit Mining Areas by Remote Sensing Techniques

Land Cover Change Detection in Urban Lake Areas Using Multi-Temporary Very High Spatial Resolution Aerial Images

Contact Info

Product

Resources

About