การประเมินผลผลิตจากการติดตามการเจริญเติบโตของข้าวนาปรังโดยค่าสัมประสิทธิ์การสะท้อนกลับจากดาวเทียมเซนติเนลวัน

Abstract

The study of "Assessing Productivity from Monitoring Growth Stage of Dry Season Rice with Backscatter Coefficient from Sentinel-1" has two objectives; 1) to monitor and find the relationship between the growths and the productivity of dry season rice with backscatter coefficient from Sentinel-1 and 2) to create the model for assessing productivity by using the backscatter coefficient according to growth stages of rice in Phichai district, Uttaradit province. The relative orbit number at 17 of Sentinel-1, GRD product, were used to apply in this study by using 16 datasets during November 2017 - May 2018. Field surveying was conducted to collect data on 34 paddy fields from farmers which is Pathum Thani 1 rice. This is for recording the boundaries, the crop start date, the crop harvest date, and yield of rice. The growth stage of rice was divided into three phases: 1) vegetative phase, 2) reproductive phase, and 3) ripening phase. The results showed that each growth phase of rice has a change in the backscatter coefficient and different scattering mechanisms and distinguished by the characteristics of the rice fields at each growth phase both VH and VV polarization. Subsequently, multiple linear regression was used to create a model for assessing dry season rice productivity. The backscatter coefficient was determined as an independent variable and the dry season rice yield was a dependent variable. The data extraction was divided into two patterns, the first pattern is the extraction of the backscatter coefficient in the whole sample rice field and the second pattern is the extraction of the backscatter coefficient at the sample point in the rice field. The most effective model from this study was the model using the maximum and mean backscatter coefficients in the whole sample rice field, VV polarization, unit is yield per rai, R Squared = 0.869, Adjusted-R Squared = 0.828, Predicted-R Squared = 0.784, and RMSE = ±40.08 kg/rai. Then, the model was used to create the productivity map of dry season rice in the other rice fields. This was performed through the process of selecting the rice fields and specifying the crop start date obtained from the study of changes in the backscatter coefficient and the normalized vegetation difference index (NDVI). The result showed that, during 2017 to 2018, 2018 to 2019, and 2019 to 2020, rice planting areas was 64,833, 61,877, and 43,895 rai, the total productivity was 48,591 46,104, 33,446 tons, and the yield per rai was 759.0, 758.6 and 767.8 kg. per rai, respectively. The rice planting areas and the total productivity are different in terms of quantity. However, the pattern of change in rice planting areas and the total productivity during the study year was consistent when compared with the dry season rice yield data from the Office of Agricultural Economics. While the yield per rai from the study of Pathum Thani 1 rice variety was similar to data from the Division of Rice Research and Development, Rice Department. This study is based on data from satellites in synthetic aperture radar (SAR) systems that have the potential to penetrate the atmosphere. It is unique in that it releases energy waves from itself, recorded data both day and night, free access, high frequency, and coverage all the world. Which is the highlight of the system for using in monitoring and evaluating rice yields. This makes it could know the yield of rice for future planning.