Articles
Longan planted area estimation and yield using remote sensing and image processing
Article number
1447_23
Pages
193 – 198
Language
English
Abstract
The aim of this paper is to estimate longan growing area and yield potential using satellite data and image processing.
Remote sensing is a technique of using satellite data of different spatial resolutions to classify objects in a measurement area.
High-profile satellite data provide better detailed data, resulting in better classification capabilities.
Satellite data are obtained by measuring the electromagnetic reflection of any object on earth with a measuring device mounted on a satellite.
Low- and medium-resolution satellite data can capture large-scale features such as agricultural fields and cultivated areas, along with their surrounding environments including water bodies and forests, while higher-resolution data provides more detailed color information.
In this research, we focused on a spatial resolution of a longan growing area in Thailand as a measured area to be determined through immediate image grading of the measuring device (IFOV). Sentinel -2 satellites carry advanced and efficient sensors consisting of two measuring instruments, the Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS). OLI measures near infrared (NIR) and short wavelength infrared (SWIR) with a resolution of 30 m and panchromatic images with a resolution of 15 m.
Emphasis was placed on the study of electromagnetic wave reflectance, indicating the amount of flowers and use that value as a factor in determining the quantity of longan flowers and analyzed together with other factors affecting the yield of longan.
Different wavelengths were selected as an indicator for longan area selection.
Red wave reflection value, green wavelength reflection value, near infrared reflection value, moisture index (MSI), infrared difference index (NDII), vegetation difference index (NDVI), moisture difference index (NDWI), and vegetation highlight index (RVI) were determined and calculated.
The actual area obtained from field survey was compared with that obtained from interpretations of the high-resolution satellite data generated in the tested area.
Using the Confusion Matrix table, it was found that the accuracy for the classification of longan plantations was 87.95% and the total accuracy was 87.42%.
Remote sensing is a technique of using satellite data of different spatial resolutions to classify objects in a measurement area.
High-profile satellite data provide better detailed data, resulting in better classification capabilities.
Satellite data are obtained by measuring the electromagnetic reflection of any object on earth with a measuring device mounted on a satellite.
Low- and medium-resolution satellite data can capture large-scale features such as agricultural fields and cultivated areas, along with their surrounding environments including water bodies and forests, while higher-resolution data provides more detailed color information.
In this research, we focused on a spatial resolution of a longan growing area in Thailand as a measured area to be determined through immediate image grading of the measuring device (IFOV). Sentinel -2 satellites carry advanced and efficient sensors consisting of two measuring instruments, the Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS). OLI measures near infrared (NIR) and short wavelength infrared (SWIR) with a resolution of 30 m and panchromatic images with a resolution of 15 m.
Emphasis was placed on the study of electromagnetic wave reflectance, indicating the amount of flowers and use that value as a factor in determining the quantity of longan flowers and analyzed together with other factors affecting the yield of longan.
Different wavelengths were selected as an indicator for longan area selection.
Red wave reflection value, green wavelength reflection value, near infrared reflection value, moisture index (MSI), infrared difference index (NDII), vegetation difference index (NDVI), moisture difference index (NDWI), and vegetation highlight index (RVI) were determined and calculated.
The actual area obtained from field survey was compared with that obtained from interpretations of the high-resolution satellite data generated in the tested area.
Using the Confusion Matrix table, it was found that the accuracy for the classification of longan plantations was 87.95% and the total accuracy was 87.42%.
Authors
R. Sukhahuta, C. Sritontip
Keywords
Dimocarpus longan, growing area estimation, yield estimation, satellite data, confusion matrix
Online Articles (24)