Comparison of maximum likelihood, support vector machines, and random forest techniques in satellite images classification

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

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

International Journal of Applied Earth Observation and Geoinformation ◽

10.1016/j.jag.2009.11.002 ◽

2010 ◽

Vol 12 ◽

pp. S27-S31 ◽

Cited By ~ 305

Author(s):

J.R. Otukei ◽

T. Blaschke

Keyword(s):

Support Vector Machines ◽

Maximum Likelihood ◽

Land Cover ◽

Decision Trees ◽

Land Cover Change ◽

Support Vector ◽

Classification Algorithms ◽

Maximum Likelihood Classification ◽

Change Assessment ◽

Vector Machines

Download Full-text

The multispectral separability of Costa Rican rainforest types with support vector machines and Random Forest decision trees

International Journal of Remote Sensing ◽

10.1080/01431160903140803 ◽

2010 ◽

Vol 31 (11) ◽

pp. 2885-2909 ◽

Cited By ~ 50

Author(s):

Steven E. Sesnie ◽

Bryan Finegan ◽

Paul E. Gessler ◽

Sirpa Thessler ◽

Zayra Ramos Bendana ◽

...

Keyword(s):

Support Vector Machines ◽

Random Forest ◽

Decision Trees ◽

Costa Rican ◽

Support Vector ◽

Vector Machines

Download Full-text

Discriminación de trigo y cebada empleando imágenes satelitales ópticas y radar

Investigaciones Geográficas Boletín del Instituto de Geografía ◽

10.14350/rig.60173 ◽

2021 ◽

Author(s):

Mario Fabián Marini

Keyword(s):

Support Vector Machines ◽

Maximum Likelihood ◽

Random Forest ◽

Vegetation Index ◽

Normalized Difference Vegetation Index ◽

Buenos Aires ◽

Support Vector ◽

Vector Machines ◽

Sentinel 2

El partido de Coronel Rosales (Buenos Aires, Argentina) se halla localizado dentro de la región pampeana austral, una de las de mayor relevancia agro productiva del país. En este contexto, el conocimiento de la superficie cultivada adquiere significativa importancia para la posterior planificación agrícola y económica. En tal sentido, la discriminación de cultivos mediante teledetección se dificulta cuando se trata de los de ciclo fenológico muy similar, como el trigo y la cebada. En este estudio se realizó una discriminación de dichos cultivos empleando imágenes de Radar de Apertura Sintética (SAR) Sentinel-1A SLC, imágenes ópticas Sentinel-2 y una combinación de ambos tipos de datos. Se incorporaron medidas de coherencia, textura e intensidad de retrodispersión extraídas de los datos SAR durante el ciclo fenológico completo. Sobre cada escena Sentinel-2 se obtuvo el Índice de Diferencia Normalizada de Vegetación (Normalized Difference Vegetation Index - NDVI). Se emplearon tres algoritmos de clasificación: Máxima Verosimilitud (Maximum Likelihood - MLC), Máquinas de Soporte Vectorial (Support Vector Machines - SVM) y Random Forest (RF). Los mejores resultados se obtuvieron al combinar imágenes ópticas y SAR empleando el clasificador RF. La combinación de las retrodispersiones VV y VH junto a la coherencia y la textura de las imágenes SAR, sumada al apilado de NDVI de imágenes ópticas, arrojó los máximos valores de precisión de la clasificación. El valor de F1 fue de 87.27% para el trigo y de 89.20% para la cebada.

Download Full-text

Assessment of Approaches for the Extraction of Building Footprints from Pléiades Images

Geomatics and Environmental Engineering ◽

10.7494/geom.2021.15.4.101 ◽

2021 ◽

Vol 15 (4) ◽

pp. 101-116

Author(s):

Lamyaa Gamal El-deen Taha ◽

Rania Elsayed Ibrahim

Keyword(s):

Support Vector Machines ◽

Maximum Likelihood ◽

High Resolution ◽

Random Forest ◽

Satellite Images ◽

Support Vector ◽

Morphological Operations ◽

Vector Machines ◽

High Resolution Images ◽

Very High

The Marina area represents an official new gateway of entry to Egypt and the development of infrastructure is proceeding rapidly in this region. The objective of this research is to obtain building data by means of automated extraction from Pléiades satellite images. This is due to the need for efficient mapping and updating of geodatabases for urban planning and touristic development. It compares the performance of random forest algorithm to other classifiers like maximum likelihood, support vector machines, and backpropagation neural networks over the well-organized buildings which appeared in the satellite images. Images were subsequently classified into two classes: buildings and non-buildings. In addition, basic morphological operations such as opening and closing were used to enhance the smoothness and connectedness of the classified imagery.The overall accuracy for random forest, maximum likelihood, support vector machines, and backpropagation were 97%, 95%, 93% and 92% respectively. It was found that random forest was the best option, followed by maximum likelihood, while the least effective was the backpropagation neural network. The completeness and correctness of the detected buildings were evaluated. Experiments confirmed that the four classification methods can effectively and accurately detect 100% of buildings from very high-resolution images. It is encouraged to use machine learning algorithms for object detection and extraction from very high-resolution images.

Download Full-text