Comparing accuracy assessments to infer superiority of image classification methods

Leeuw, Jan de; Jia, Hui-Xing; Yang, Lei; Liu, X.; Schmidt, K.S.; Skidmore, Andrew K.

doi:10.1080/01431160500275762

Cited by 134 publications

(72 citation statements)

References 8 publications

(11 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, this may not be appropriate if the same sample of sites is used in the comparison [29][30][31], as these coefficients assume that the samples used in their calculations are independent. We have used the same set of reference points for testing the accuracy of maps produced by MLC and PCC methods for the same year, to avoid the difference of accuracy due to sampling variability.…”

Section: Comparing Classifier Performancementioning

confidence: 99%

“…We have used the same set of reference points for testing the accuracy of maps produced by MLC and PCC methods for the same year, to avoid the difference of accuracy due to sampling variability. For this reason, we have performed McNemar's test [30] to evaluate the superiority of the LULC maps resulting from post-classification over the MLC classified maps. The McNemar's test is preferable because it is parametric, very simple to understand and execute.…”

Section: Comparing Classifier Performancementioning

confidence: 99%

See 1 more Smart Citation

Improving the Accuracy of Land Use and Land Cover Classification of Landsat Data Using Post-Classification Enhancement

2009

View full text Add to dashboard Cite

Classifying remote sensing imageries to obtain reliable and accurate land use and land cover (LULC) information still remains a challenge that depends on many factors such as complexity of landscape, the remote sensing data selected, image processing and classification methods, etc. The aim of this paper is to extract reliable LULC information from Landsat imageries of the Lower Hunter region of New South Wales, Australia. The classical maximum likelihood classifier (MLC) was first applied to classify Landsat-MSS of 1985 and Landsat-TM of 1995 and 2005. The major LULC identified were Woodland, Pasture/scrubland, Vineyard, Built-up and Water-body. By applying post-classification correction (PCC) using ancillary data and knowledge-based logic rules the overall classification accuracy was improved from about 72% to 91% for 1985 map, 76% to 90% for 1995 map and 79% to 87% for 2005 map. The improved overall Kappa statistics due to PCC were 0.88 for the 1985 map, 0.86 for 1995 and 0.83 for 2005. The PCC maps, assessed by McNemar's test, were found to have much higher accuracy in comparison to their counterpart MLC maps. The overall improvement in classification accuracy of the LULC maps is significant in terms of their potential use for land change modelling of the region.

show abstract

Section: Comparing Classifier Performancementioning

confidence: 99%

Section: Comparing Classifier Performancementioning

confidence: 99%

Improving the Accuracy of Land Use and Land Cover Classification of Landsat Data Using Post-Classification Enhancement

2009

View full text Add to dashboard Cite

show abstract

“…Accuracy and comparison were determined by overall accuracy coefficient and Kappa index. Significance from Kappa indexes was analyzed through Z testing, which enabled to verify if classification in the neural networks was considered better than a random classification (De Leeuw et al, 2006). The selected neural network was applied to fruits in the 10 maturity weeks evaluated, and the percentage of classified fruits pertaining to Immature class (A) and Matured class (B).…”

Section: Fruit Classification To Determine Harvest Moment In Functionmentioning

confidence: 99%

Determination of macaw fruit harvest period by biospeckle laser technique

Costa¹,

Pinto²,

JÃonior³

et al. 2017

Afr. J. Agric. Res.

View full text Add to dashboard Cite

Macaw palm has been stood out as a raw material for the production of bioenergy, because it has high productivity of oil and less emission of polluting waste during combustion, meeting the worldwide demand for sustainable energy sources. The aims of this research were the evaluation of response of the biologic activity measured by the optical technique of the biospeckle laser applied to macaw palm fruits at different maturity weeks and develop a classifier in function of biologic activity to determine the harvest period related with oil content in the fruits. To perform the experiment, 10 weeks fruits different maturity stages were evaluated. The biospeckle laser images were obtained by illuminating the epicarp of each fruit. The biological activity was quantified by absolute value of difference algorithm applied to biospeckle images. A neural network was developed to classify the fruits which were closer to harvest in function of biologic activity. Biologic activity showed a significant linear ratio (R 2 = 0.913) with the maturation of fruits. Classification results have shown that fruits from 59 th week after flowering are ideal for harvest and present the highest oil levels.

show abstract

“…McNemar's test allows separate classifications that were assessed using the same set of accuracy points to be compared without bias. McNemar's test is a non-parametric test based on a binary distinction between correct and incorrect class allocations that uses a 2 by 2 matrix to calculate a chi squared value [35,36]. McNemar's test of the multispectral NAIP, NDVI, and CHM classification compared to the multispectral NAIP and NDVI classification shows that the classifications are statistically different at the p = 0.01 level.…”

Section: Significance Of Including Lidar Datamentioning

confidence: 99%

Fusion of High Resolution Aerial Multispectral and LiDAR Data: Land Cover in the Context of Urban Mosquito Habitat

2011

View full text Add to dashboard Cite

Abstract:Remotely sensed multi-spectral and -spatial data facilitates the study of mosquito-borne disease vectors and their response to land use and cover composition in the urban environment. In this study we assess the feasibility of integrating remotely sensed multispectral reflectance data and LiDAR (Light Detection and Ranging)-derived height information to improve land use and land cover classification. Classification and Regression Tree (CART) analyses were used to compare and contrast the enhancements and accuracy of the multi-sensor urban land cover classifications. Eight urban land-cover classes were developed for the city of Tucson, Arizona, USA. These land cover classes focus on pervious and impervious surfaces and microclimate landscape attributes that impact mosquito habitat such as water ponds, residential structures, irrigated lawns, shrubs and trees, shade, and humidity. Results show that synergistic use of LiDAR, multispectral and the Normalized Difference Vegetation Index data produced the most accurate urban land cover classification with a Kappa value of 0.88. Fusion of multi-sensor data leads to a better land cover product that is suitable for a variety of urban applications such as exploring the relationship between neighborhood composition and adult mosquito abundance data to inform public health issues.

show abstract

Comparing accuracy assessments to infer superiority of image classification methods

Cited by 134 publications

References 8 publications

Improving the Accuracy of Land Use and Land Cover Classification of Landsat Data Using Post-Classification Enhancement

Improving the Accuracy of Land Use and Land Cover Classification of Landsat Data Using Post-Classification Enhancement

Determination of macaw fruit harvest period by biospeckle laser technique

Fusion of High Resolution Aerial Multispectral and LiDAR Data: Land Cover in the Context of Urban Mosquito Habitat

Contact Info

Product

Resources

About