Land cover change assessment using decision trees, support vector machines and maximum likelihood classification algorithms

Context: Nowadays, the images of the Earth surface and the algorithms for their classification are widely available. In particular, the algorithms are promising in the differentiating of cotton crops stages, but it is necessary to establish the capabilities of the different algorithms in order to identify their advantages, and disadvantages. Method: This paper describes the assessment process in which the Support Vector Machines (SVM) and random-forest technique (decision trees) are compared with the maximum likelihood estimation when differentiating the stages of cotton crops. A RapidEye satellite image of a geographic area in the municipality of San Pelayo, Cordoba (Colombia), is used for the study. Using a set of sampling polygons, a random sample of 6000 pixels was taken (2000 training and 4000 for validating the classifications.) Confusion matrices, and R (data processing and analysis software) were used during the validation process Results: The maximun likelihood estimation presented a correct classification percentage of 68.95%. SVM correctly classified 81.325% of the cases and the decision trees correctly classified 78.925%. The confidence test for the classifications showed non-overlapping intervals, and SVM obtained the highest values. Conclusions: It was possible to confirm the superiority of the technique based on support vector machines for the proposed verification zones. However, this technique requires a number of classes that comprehensively represent the variations of the image (in order to guarantee a minimum number of support vectors) to avoid confusion in the classification of non-sampled areas. This was less evident in the other two classification techniques analysed.

Download Full-text

Efficient and Private Scoring of Decision Trees, based on Pre-Computation Technique with Support Vector Machines and Logistic Regression Model

International Journal of Computing Communications and Networking ◽

10.30534/ijccn/2018/16722018 ◽

2018 ◽

Vol 7 (2) ◽

pp. 96-99

Author(s):

Keyword(s):

Logistic Regression ◽

Support Vector Machines ◽

Regression Model ◽

Decision Trees ◽

Logistic Regression Model ◽

Support Vector ◽

Computation Technique ◽

Vector Machines

Download Full-text

Quantifying Land Cover Changes in a Mediterranean Environment Using Landsat TM and Support Vector Machines

Forests ◽

10.3390/f11070750 ◽

2020 ◽

Vol 11 (7) ◽

pp. 750 ◽

Cited By ~ 2

Author(s):

Sotiria Fragou ◽

Kleomenis Kalogeropoulos ◽

Nikolaos Stathopoulos ◽

Panagiota Louka ◽

Prashant K. Srivastava ◽

...

Keyword(s):

Support Vector Machines ◽

Land Cover ◽

Forest Cover ◽

Spectral Band ◽

Mixed Forest ◽

Landsat Tm ◽

Geographical Information ◽

Support Vector ◽

Geoinformation Technologies ◽

Vector Machines

The rapid advent in geoinformation technologies, such as Earth Observation (EO) and Geographical Information Systems (GIS), has made it possible to observe and monitor the Earth’s environment on variable geographical scales and analyze those changes in both time and space. This study explores the synergistic use of Landsat EO imagery and Support Vector Machines (SVMs) in obtaining Land Use/Land Cover (LULC) mapping and quantifying its spatio-temporal changes for the municipality of Mandra–Idyllia, Attica Region, Greece. The study area is representative of typical Mediterranean landscape in terms of physical structure and coverage of species composition. Landsat TM (Thematic Mapper) images from 1993, 2001 and 2010 were acquired, pre-processed and classified using the SVMs classifier. A total of nine basic classes were established. Eight spectral band ratios were created in order to incorporate them in the initial variables of the image. For validating the classification, in-situ data were collected for each LULC type during several field surveys that were conducted in the area. The overall classification accuracy for 1993, 2001 and 2010 Landsat images was reported as 89.85%, 91.01% and 90.24%, respectively, and with a statistical factor (K) of 0.96, 0.89 and 0.99, respectively. The classification results showed that the total extent of forests within the studied period represents the predominant LULC, despite the intense human presence and its impacts. A marginal change happened in the forest cover from 1993 to 2010, although mixed forest decreased significantly during the studied period. This information is very important for future management of the natural resources in the studied area and for understanding the pressures of the anthropogenic activities on the natural environment. All in all, the present study demonstrated the considerable promise towards the support of geoinformation technologies in sustainable environmental development and prudent resource management.

Download Full-text