A systematic comparison of different object-based classification techniques using high spatial resolution imagery in agricultural environments

Li, Manchun; Ma, Lei; Blaschke, Thomas; Cheng, Liang; Tiede, Dirk

doi:10.1016/j.jag.2016.01.011

Cited by 168 publications

(126 citation statements)

References 43 publications

Supporting

Mentioning

118

Contrasting

Unclassified

Order By: Relevance

“…Likewise, the SVM algorithm was stable with regard to scale and relatively robust in the selection of object features and could perform better with limited training samples than other algorithms [21,50,51]. In this study, the performance of SVM was most stable in different image analysis approaches, but could not only well extract mature and pure mangrove species (SA1), but could also extract partially mixed mangrove species (HT) due to its ability to locate an optimal separating hyperplane for high dimensional features.…”

Section: The Comparison and Selection Of Machine Learning Algorithmsmentioning

confidence: 99%

“…For the C, we tested the following values: 2, 50, 100, 200, 300, 400, and 500; for the g, 0.25, 0.1, 0.01, 0.001, 0.0001, and 0.00001 were examined. In pixel-based SVM classification, a two-dimensional grid search with internal validation was employed to search optimal values of C and g [21]. A total of ten values for the parameter C (1, 2, 4, 8, 16, 32, 64, 128, 256, and 512) and a total of eight values for the g (0.00001, 0.0001, 0.001, 0.01, 0.1, 1, 10, and 100) were tested in the pixel-based SVM classification.…”

Section: Tuning Of Machine Learning Algorithm Parametersmentioning

confidence: 99%

“…Therefore, a statistical comparison using the Kappa coefficient was unwarranted [20,39]. As a result, the Monte Carlo test [20], McNemar's test [40], Kruskal-Wallis test [21], Nemenyi post-hoc test [21] et al statistical test methods were recommended for assessing whether there was a statistically significant difference between the classification methods. McNemar's test is non-parametric and based on the error matrices of the classifier, determining if column and row frequencies are equal (null hypothesis) [40,41].…”

Section: Accuracy Assessmentmentioning

confidence: 99%

“…In general, the DT algorithm was sensitive to segmentation scale variation and robust to the selection of object features, and could perform well at an optimal segmentation scale [21]. For artificial mangrove species discrimination, DT showed an obvious difference between pixel-based and object-based image analysis, and performed poorly in the face of complex mangrove species.…”

Section: The Comparison and Selection Of Machine Learning Algorithmsmentioning

confidence: 99%

“…As a classic machine learning algorithm, decision tree (DT), referred to as CART (classification and regression tree) or C4.5 algorithm et al, is less used in mangrove classification. RF is a relatively new technique for mangrove species mapping, though it has been widely applied in land [20,21] and plant species [8,22] classification with different sensors in recent years. In addition, DT, RF and SVM all can be implemented in eCognition 9.0 (a popular object-based image analysis software, Trimble, Sunnyvale, CO, USA) with streamlined production models.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Artificial Mangrove Species Mapping Using Pléiades-1: An Evaluation of Pixel-Based and Object-Based Classifications with Selected Machine Learning Algorithms

et al. 2018

View full text Add to dashboard Cite

Abstract:In the dwindling natural mangrove today, mangrove reforestation projects are conducted worldwide to prevent further losses. Due to monoculture and the low survival rate of artificial mangroves, it is necessary to pay attention to mapping and monitoring them dynamically. Remote sensing techniques have been widely used to map mangrove forests due to their capacity for large-scale, accurate, efficient, and repetitive monitoring. This study evaluated the capability of a 0.5-m Pléiades-1 in classifying artificial mangrove species using both pixel-based and object-based classification schemes. For comparison, three machine learning algorithms-decision tree (DT), support vector machine (SVM), and random forest (RF)-were used as the classifiers in the pixel-based and object-based classification procedure. The results showed that both the pixel-based and object-based approaches could recognize the major discriminations between the four major artificial mangrove species. However, the object-based method had a better overall accuracy than the pixel-based method on average. For pixel-based image analysis, SVM produced the highest overall accuracy (79.63%); for object-based image analysis, RF could achieve the highest overall accuracy (82.40%), and it was also the best machine learning algorithm for classifying artificial mangroves. The patches produced by object-based image analysis approaches presented a more generalized appearance and could contiguously depict mangrove species communities. When the same machine learning algorithms were compared by McNemar's test, a statistically significant difference in overall classification accuracy between the pixel-based and object-based classifications only existed in the RF algorithm. Regarding species, monoculture and dominant mangrove species Sonneratia apetala group 1 (SA1) as well as partly mixed and regular shape mangrove species Hibiscus tiliaceus (HT) could well be identified. However, for complex and easily-confused mangrove species Sonneratia apetala group 2 (SA2) and other occasionally presented mangroves species (OT), only major distributions could be extracted, with an accuracy of about two-thirds. This study demonstrated that more than 80% of artificial mangroves species distribution could be mapped.

show abstract

Section: The Comparison and Selection Of Machine Learning Algorithmsmentioning

confidence: 99%

Section: Tuning Of Machine Learning Algorithm Parametersmentioning

confidence: 99%

Section: Accuracy Assessmentmentioning

confidence: 99%

Section: The Comparison and Selection Of Machine Learning Algorithmsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Artificial Mangrove Species Mapping Using Pléiades-1: An Evaluation of Pixel-Based and Object-Based Classifications with Selected Machine Learning Algorithms

et al. 2018

View full text Add to dashboard Cite

show abstract

Quality Assessment of Worldview-3 Stereo Imagery Derived Models Over Millennial Olive Groves

Šiljeg

Domazetović

Marić

et al. 2021

Communications in Computer and Information Science

View full text Add to dashboard Cite

Optimized Rule Sets for Automatic Landslide Characteristic Detection in a Highly Vegetated Forests

Pradhan

Mezaal

2017

Laser Scanning Applications in Landslide Assessment

View full text Add to dashboard Cite

A systematic comparison of different object-based classification techniques using high spatial resolution imagery in agricultural environments

Cited by 168 publications

References 43 publications

Artificial Mangrove Species Mapping Using Pléiades-1: An Evaluation of Pixel-Based and Object-Based Classifications with Selected Machine Learning Algorithms

Artificial Mangrove Species Mapping Using Pléiades-1: An Evaluation of Pixel-Based and Object-Based Classifications with Selected Machine Learning Algorithms

Quality Assessment of Worldview-3 Stereo Imagery Derived Models Over Millennial Olive Groves

Optimized Rule Sets for Automatic Landslide Characteristic Detection in a Highly Vegetated Forests

Contact Info

Product

Resources

About