Comparative Analysis of Pixel-Based and Object-Based Classification of High Resolution Remote Sensing Images – A Review

Aggarwal, Nikita; Srivastava, Mohit; Dutta, Maitreyee

doi:10.14445/22315381/ijett-v38p202

Cited by 27 publications

(16 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite its broad application, kappa statistics can be very misleading to assess or communicate the accuracy of classification due to its high dependency on the variation of class prevalence [42]. An automatic comparison between PB and OO methods in LULC classification, based on confusion matrices, could result very useful to define the best approach in terms of accuracy of the various LULC classes [43] and to assess the improvement achieved with the OO approach, since the final accuracy is always conditioned by the proprieties of the input data [44], the quality of training information, and the peculiarities of the study areas.…”

Section: Introductionmentioning

confidence: 99%

Object-Oriented LULC Classification in Google Earth Engine Combining SNIC, GLCM, and Machine Learning Algorithms

2020

View full text Add to dashboard Cite

Google Earth Engine (GEE) is a versatile cloud platform in which pixel-based (PB) and object-oriented (OO) Land Use–Land Cover (LULC) classification approaches can be implemented, thanks to the availability of the many state-of-art functions comprising various Machine Learning (ML) algorithms. OO approaches, including both object segmentation and object textural analysis, are still not common in the GEE environment, probably due to the difficulties existing in concatenating the proper functions, and in tuning the various parameters to overcome the GEE computational limits. In this context, this work is aimed at developing and testing an OO classification approach combining the Simple Non-Iterative Clustering (SNIC) algorithm to identify spatial clusters, the Gray-Level Co-occurrence Matrix (GLCM) to calculate cluster textural indices, and two ML algorithms (Random Forest (RF) or Support Vector Machine (SVM)) to perform the final classification. A Principal Components Analysis (PCA) is applied to the main seven GLCM indices to synthesize in one band the textural information used for the OO classification. The proposed approach is implemented in a user-friendly, freely available GEE code useful to perform the OO classification, tuning various parameters (e.g., choose the input bands, select the classification algorithm, test various segmentation scales) and compare it with a PB approach. The accuracy of OO and PB classifications can be assessed both visually and through two confusion matrices that can be used to calculate the relevant statistics (producer’s, user’s, overall accuracy (OA)). The proposed methodology was broadly tested in a 154 km2 study area, located in the Lake Trasimeno area (central Italy), using Landsat 8 (L8), Sentinel 2 (S2), and PlanetScope (PS) data. The area was selected considering its complex LULC mosaic mainly composed of artificial surfaces, annual and permanent crops, small lakes, and wooded areas. In the study area, the various tests produced interesting results on the different datasets (OA: PB RF (L8 = 72.7%, S2 = 82%, PS = 74.2), PB SVM (L8 = 79.1%, S2 = 80.2%, PS = 74.8%), OO RF (L8 = 64%, S2 = 89.3%, PS = 77.9), OO SVM (L8 = 70.4, S2 = 86.9%, PS = 73.9)). The broad code application demonstrated very good reliability of the whole process, even though the OO classification process resulted, sometimes, too demanding on higher resolution data, considering the available computational GEE resources.

show abstract

Section: Introductionmentioning

confidence: 99%

Object-Oriented LULC Classification in Google Earth Engine Combining SNIC, GLCM, and Machine Learning Algorithms

2020

View full text Add to dashboard Cite

show abstract

“…Besides the increasing precision, the feature extraction procedure also enhances the computational speed of the classifier. There are certain steps that are required to be followed in order to perform the feature extraction from an object as defined below [14]. a.…”

Section: Feature Extractionmentioning

confidence: 99%

“…The result has been calculated according to the kappa coefficient and the accuracy. Nikita Aggarwal et al, (2016) [14] Pixel-Based Classificatio n (PBC) and Object-Based Classificatio n (OBC)…”

Section: Lu and Q We (2007) [3]mentioning

confidence: 99%

A review of remotely sensed satellite image classification

Dhingra

Kumar

2019

IJECE

View full text Add to dashboard Cite

Satellite image classification has a vital role for the extraction and analysis of the useful satellite image information. This paper comprises the study of the satellite images classification and Remote Sensing along with a brief overview of the previous studies that are proposed in this field. In this paper, the existing work has been explained utilizing the classification techniques on satellite images of Alwar region in India that covers decent land cover features like Vegetation, Water, Urban, Barren, and Rocky regions. The post- implementation of the classification algorithms, the classified image is obtained displaying different classes that are represented by different colours. Each feature is represented by a different colour and can be easily perceived from the image obtained after classification. The focus of this study is on enhancing the classification accuracy by using proper classifiers along with the novel feature extraction techniques and pre-processing steps. Work of different authors is being discussed in a tabular form defining the methods and outcomes of the respective studies.

show abstract

“…Pixel-based algorithms are theoretically simple and known to well perform for low and medium-resolution images (Hussain et al, 2013). However, in the high-resolution images, pixels are not spatially independent, and it makes these conventional pixel-based methods less effective than objectbased methods (Aggarwal et al, 2016). Compared with pixel-based approaches, object-based image analysis assumes that the landscapes are composed of meaningful objects corresponding to ground entities and patches of surface cover (Blaschke et al, 2000).…”

Section: Introductionmentioning

confidence: 99%

Wildfire Damage Assessment Using Multi-Temporal Sentinel-2 Data

Chung

Jung

Kim

2019

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. Recently, the drastic climate changes have increased the importance of wildfire monitoring and damage assessment as well as the possibility of wildfire occurrence. Estimation of wildfire damage provides the information on wildfire-induced ecological changes and supports the decision-making process for post-fire treatment activities. For accurate wildfire damage assessment, the discrimination between disaster-induced and natural changes is crucial because they usually coupled together. In this study, Sentinel-2 images were employed to assess the damage from a wildfire, which occurred in the coniferous forest of Gangneung, Gangwon Province, South Korea on April 2019. The images were captured from both Sentinel-2A and -2B, shortening the temporal interval of available pre- and post-fire images. Multi-temporal image analysis was performed in both object and pixel-based with two commonly used spectral indices, NDVI and NBR. Additional image pair from the same period of 2018 was used to distinguish the fire-affected regions from the naturally changed area and compared with the results from using only one pair of images from 2019. The experimental results showed that the change detection performance could be affected by the number of image pairs and spectral indices used to discriminate burned region from unburned region. Thus it verified the significance of adequately employing annual multi-pair satellite images for wildfire damage assessment.

show abstract

Comparative Analysis of Pixel-Based and Object-Based Classification of High Resolution Remote Sensing Images – A Review

Cited by 27 publications

References 0 publications

Object-Oriented LULC Classification in Google Earth Engine Combining SNIC, GLCM, and Machine Learning Algorithms

Object-Oriented LULC Classification in Google Earth Engine Combining SNIC, GLCM, and Machine Learning Algorithms

A review of remotely sensed satellite image classification

Wildfire Damage Assessment Using Multi-Temporal Sentinel-2 Data

Contact Info

Product

Resources

About