Yield Prediction for Winter Wheat with Machine Learning Models Using Sentinel-1, Topography, and Weather Data
We train and compare the performance of two different machine learning algorithms to learn changes in winter wheat production for fields from the southwest of Sweden. As input to these algorithms, we use cloud-penetrating Sentinel-1 polarimetry radar data together with respective field topography an...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | Agriculture |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2077-0472/13/4/813 |
| _version_ | 1797606871802052608 |
|---|---|
| author | Oliver Persson Bogdanovski Christoffer Svenningsson Simon Månsson Andreas Oxenstierna Alexandros Sopasakis |
| author_facet | Oliver Persson Bogdanovski Christoffer Svenningsson Simon Månsson Andreas Oxenstierna Alexandros Sopasakis |
| author_sort | Oliver Persson Bogdanovski |
| collection | DOAJ |
| description | We train and compare the performance of two different machine learning algorithms to learn changes in winter wheat production for fields from the southwest of Sweden. As input to these algorithms, we use cloud-penetrating Sentinel-1 polarimetry radar data together with respective field topography and local weather over four different years. We note that all of the input data were freely available. During training, we used information on winter wheat production over the fields of interest which was available from participating farmers. The two machine learning models we trained are the Light Gradient-Boosting Machine and a Feed-forward Neural Network. Our results show that Sentinel-1 data contain valuable information which can be used for training to predict winter wheat yield once two important steps are taken: performing a critical transformation of each pixel in the images to align it to the training data and then following up with despeckling treatment. Using this approach, we were able to achieve a top root mean square error of 0.75 tons per hectare and a top accuracy of 86% using a k-fold method with <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mi>k</mi><mo>=</mo><mn>5</mn></mrow></semantics></math></inline-formula>. More importantly, however, we established that Sentinel-1 data alone are sufficient to predict yield with an average root mean square error of 0.89 tons per hectare, making this method feasible to employ worldwide. |
| first_indexed | 2024-03-11T05:21:05Z |
| format | Article |
| id | doaj.art-9e4a1d809bee49f694cb0d5922d6779f |
| institution | Directory Open Access Journal |
| issn | 2077-0472 |
| language | English |
| last_indexed | 2024-03-11T05:21:05Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Agriculture |
| spelling | doaj.art-9e4a1d809bee49f694cb0d5922d6779f2023-11-17T17:53:53ZengMDPI AGAgriculture2077-04722023-03-0113481310.3390/agriculture13040813Yield Prediction for Winter Wheat with Machine Learning Models Using Sentinel-1, Topography, and Weather DataOliver Persson Bogdanovski0Christoffer Svenningsson1Simon Månsson2Andreas Oxenstierna3Alexandros Sopasakis4Department of Mathematics, Faculty of Science, Lund University, 221 00 Lund, SwedenDepartment of Mathematics, Faculty of Science, Lund University, 221 00 Lund, SwedenNiftitech AB, Hedvig Möllers gata 12, 223 55 Lund, SwedenT-Kartor AB, Olof Mohlins väg 12, 291 62 Kristianstad, SwedenDepartment of Mathematics, Faculty of Science, Lund University, 221 00 Lund, SwedenWe train and compare the performance of two different machine learning algorithms to learn changes in winter wheat production for fields from the southwest of Sweden. As input to these algorithms, we use cloud-penetrating Sentinel-1 polarimetry radar data together with respective field topography and local weather over four different years. We note that all of the input data were freely available. During training, we used information on winter wheat production over the fields of interest which was available from participating farmers. The two machine learning models we trained are the Light Gradient-Boosting Machine and a Feed-forward Neural Network. Our results show that Sentinel-1 data contain valuable information which can be used for training to predict winter wheat yield once two important steps are taken: performing a critical transformation of each pixel in the images to align it to the training data and then following up with despeckling treatment. Using this approach, we were able to achieve a top root mean square error of 0.75 tons per hectare and a top accuracy of 86% using a k-fold method with <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mi>k</mi><mo>=</mo><mn>5</mn></mrow></semantics></math></inline-formula>. More importantly, however, we established that Sentinel-1 data alone are sufficient to predict yield with an average root mean square error of 0.89 tons per hectare, making this method feasible to employ worldwide.https://www.mdpi.com/2077-0472/13/4/813precision agricultureSentinel-1 SARmachine learningyield predictiondespeckling |
| spellingShingle | Oliver Persson Bogdanovski Christoffer Svenningsson Simon Månsson Andreas Oxenstierna Alexandros Sopasakis Yield Prediction for Winter Wheat with Machine Learning Models Using Sentinel-1, Topography, and Weather Data Agriculture precision agriculture Sentinel-1 SAR machine learning yield prediction despeckling |
| title | Yield Prediction for Winter Wheat with Machine Learning Models Using Sentinel-1, Topography, and Weather Data |
| title_full | Yield Prediction for Winter Wheat with Machine Learning Models Using Sentinel-1, Topography, and Weather Data |
| title_fullStr | Yield Prediction for Winter Wheat with Machine Learning Models Using Sentinel-1, Topography, and Weather Data |
| title_full_unstemmed | Yield Prediction for Winter Wheat with Machine Learning Models Using Sentinel-1, Topography, and Weather Data |
| title_short | Yield Prediction for Winter Wheat with Machine Learning Models Using Sentinel-1, Topography, and Weather Data |
| title_sort | yield prediction for winter wheat with machine learning models using sentinel 1 topography and weather data |
| topic | precision agriculture Sentinel-1 SAR machine learning yield prediction despeckling |
| url | https://www.mdpi.com/2077-0472/13/4/813 |
| work_keys_str_mv | AT oliverperssonbogdanovski yieldpredictionforwinterwheatwithmachinelearningmodelsusingsentinel1topographyandweatherdata AT christoffersvenningsson yieldpredictionforwinterwheatwithmachinelearningmodelsusingsentinel1topographyandweatherdata AT simonmansson yieldpredictionforwinterwheatwithmachinelearningmodelsusingsentinel1topographyandweatherdata AT andreasoxenstierna yieldpredictionforwinterwheatwithmachinelearningmodelsusingsentinel1topographyandweatherdata AT alexandrossopasakis yieldpredictionforwinterwheatwithmachinelearningmodelsusingsentinel1topographyandweatherdata |