Crop classification and mapping using remote sensing and machine learning

Research results on crop classification and mapping using machine learning (ML) with remotely sensed data are presented. Studies were conducted in 2022, 2023 on the land use of the farms located in the Novosibirsk region. Sentinel-2 and Landsat 8-9 images were used. The features for training the ML models were six spectral bands and five vegetation indices for different dates of the growing seasons. XGBoost, KNN, RF and SVM supervised learning ML algorithms and deep-FNN neural network were applied. The XGBoost, KNN and RF models showed high classification accuracies of 93–97% at 30 m/pixel resolution and 80-90% at 90 m/pixel resolution. The deep-FNN model showed the lowest results with an accuracy of 78 to 92% at 30 m/pixel resolution. The overall 8–12% reduction in accuracy at 90 m/pixel resolution compared to 30 m/pixel resolution emphasizes the importance of scale for effective crop recognition. Also, training the models on the combined Sentinel-2 and Landsat 8-9 satellite data at 30 m/pixel resolution yielded higher values of the F1-score metric than on the data separately for each of these satellites. Various evaluation metrics (F1-score and ROC-AUCscore) confirmed that the XGBoost model was the best performing and most accurate. The best overall classification was achieved for corn, barley, annual and perennial grasses, and fallow, with some decrease in accuracy for oats, peas, vetch and soft winter wheat. The lowest accuracy was observed in the classification of potatoes, spring rape, soft spring wheat and fallow. The results of the research emphasize the importance of ML model selection and satellite imagery resolution scale for successful crop classification.

1. Plotnikov D.E., Hvostikov S.A., Bartalev S.A. Method for automated crop types mapping using remote sensing data and a plant growth simulation model. Sovremenny`e problemy` distancionnogo zondirovaniya Zemli iz kosmosa = Current problems in remote sensing of the Earth from space, 2018, vol. 15, no. 4. pp. 131–141. (In Russian).

2. Yakushev V.P., Zaxaryan Yu.G., Bloxina S.Yu. Current problems and prospects for the use of remote sensing of the Earth in agriculture. Sovremenny`e problemy` distancionnogo zondirovaniya Zemli iz kosmosa = Current problems in remote sensing of the Earth from space, 2022, vol. 19. no. 1. pp. 287–294. (In Russian). DOI: 10.21046/2070-7401-2022-19-1-287-294.

3. Illarionova L.V., Stepanov A.S., Fomina E.A. Recognition and classification of agricultural crops in Khabarovsk region using NDVI and EVI. Sovremenny`e problemy` distancionnogo zondirovaniya Zemli iz kosmosa = Current problems in remote sensing of the Earth from space, 2023, vol. 20, no. 2, p. 155. (In Russian). DOI: 10.21046/2070-7401-2023-20-2-155-165.

4. Kononov V.M., Asadullaev R.G., Kuz`menko N.I. Algorithm of multi-spectral satellite data preparation for agricultural crop classification. Nauchny`j rezul`tat. Informacionny`e texnologii = Research result. Information technologies, 2020, vol. 5, no. 2, pp. 18–24. (In Russian). DOI: 10.18413/2518-1092-2020-5-2-0-3.

5. Wu B., Zhang M., Zeng H., Tian F., Potgieter A.B., Qin X., Yan N., Chang S., Zhao Y., Dong Q., Boken V., Plotnikov D., Guo H., Wu F., Zhao H., Deronde B., Tits L., Loupian E. Challenges and opportunities in remote sensing-based crop monitoring: a review. National Science Review, 2023, vol. 10, no. 4, р. 290. DOI: 10.1093/nsr/nwac290.

6. Weiss M., Jacob F., Duveiller G. Remote sensing for agricultural applications: A meta-review. Remote sensing of environment, 2020, vol. 236, p. 111402. DOI: 10.1016/j.rse.2019.111402.

7. Zeng Y., Hao D., Huete A., Dechant B., Berry J., Chen J.M., Joiner J., Frankenberg C., Bond-Lamberty B., Ryu Y., Xiao J., Asrar G.R., Chen M. Optical vegetation indices for monitoring terrestrial ecosystems globally. Nature Reviews Earth & Environment, 2022, vol. 3, no. 7, р. 477–493. DOI: 10.1038/s43017-022-00298-5.

8. Chakhar A., Ortega-Terol D., Hernández-López D., Ballesteros R., Ortega J.F., Moreno M.A. Assessing the accuracy of multiple classification algorithms for crop classification using Landsat-8 and Sentinel-2 data. Remote sensing, 2020, vol. 12, no. 11, p. 1735. DOI: 10.3390/rs12111735.

9. Song X.P., Huang W., Hansen M.C., Potapov P. An evaluation of Landsat, Sentinel-2, Sentinel-1 and MODIS data for crop type mapping. Science of Remote Sensing, 2021, vol. 3, p. 100018. DOI: 10.1016/j.srs.2021.100018.

10. Blickensdörfer L., Schwieder M., Pflugmacher D., Nendel C., Erasmi S., Hostert P. Mapping of crop types and crop sequences with combined time series of Sentinel-1, Sentinel-2 and Landsat 8 data for Germany. Remote sensing of environment, 2022, vol. 269, p. 112831. DOI: 10.1016/j.rse.2021.112831.

11. Proxorecz I.O., Stepanov A.S. Mapping of the Khabarovsk region arable lands by machine learning using Sentinel-2 images. Sistemy` i sredstva informatiki = Systems and means of informatics, 2024, vol. 34, no. 1, pp. 57–69. (In Rusian). DOI: 10.14357/08696527240105.

12. Dubrovin K.N., Stepanov A.S., VerxoturovA.L., Aseeva T.A. Crop identification using radar images. Informatika i avtomatizaciya = Informatics and Automation, 2022, vol. 21, no. 2, pp. 405–426. (In Russian). DOI: 10.15622/ia.21.2.7.

13. Guo L., Zhao S., Gao J., Zhang H., Zou Y., Xiao X. A novel workflow for crop type mapping with a time series of synthetic aperture radar and optical images in the Google Earth Engine. Remote Sensing, 2022, vol. 14, no. 21, p. 5458 DOI: 10.3390/rs14215458.

14. Faqe I.G.R., Rasul A., Abdulla H. Improving Crop Classification Ac-curacy with Integrated Sentinel-1 and Sentinel-2 Data: a Case Study of Barley and Wheat. Journal of Geovisualization and Spatial Analysi, 2023, vol. 7, no. 2, p. 22. DOI: 10.1007/s41651-023-00152-2.

15. Vorob`ev N.I., Puxal`skij Ya.V., Astapova M.A., Surin V.G., Pishhik V.N. Digital processing of photometric data of remote sensing of winter rye fields. Agrarny`j vestnik Urala = Agrarian Bulletin of the Urals, 2024, vol. 24, no. 2, pp. 152– 162. (In Russian). DOI: 10.32417/1997-4868-2024-24-02-152-162.

16. Asadullaev R.G., Kuz`menko N.I. Technology of intelligent agricultural crops recognition by neural network based on multispectral multitemporal Earth remote sensing data. E`konomika. Informatika = Economics. Information technologies, 2022, vol. 49, no. 1, pp. 159–168. (In Russian). DOI: 10.52575/2687-0932-2022-49-1-159-168.

17. Joshi A., Pradhan B., Gite S., Chakraborty S. Remote-sensing data and deep-learning techniques in crop mapping and yield prediction: A systematic review. Remote Sensing, 2023, vol. 15, no. 8, p. 2014. DOI: 10.3390/rs15082014.

18. Tokarev K.E., Lebed` N.I. Multiclass recognition of agricultural crops by a recurrent deep learning neural network with convolutional layers based on high-resolution color aerial photographs. Mezhdunarodny`j sel`skoxozyajstvenny- `j zhurnal = International agricultural journal, 2024, no. 2, pp. 192–195. (In Russian). DOI: 10.55186/25876740_2024_67_2_192.

19. Alami M.M., Mansouri E., Imani Y., Bourja O., Lahlou O., Zennayi Y., Bourzeixet F., Houmma I.H., Hadria R. Crop mapping using supervised machine learning and deep learning: A systematic literature review. International Journal of Remote Sensing, 2023, vol. 44, no. 8, pp. 2717– 2753. DOI: 10.1080/01431161.2023.2205984.

20. Orynbaikyzy A., Gessner U., Mack B., Conrad C. Crop Type Classification Using Fusion of Sentinel-1 and Sentinel-2 Data: Assessing the Impact of Feature Selection, Optical Data Availability, and Parcel Sizes on the Accuracies. Remote Sensing, 2020, vol. 12, no. 17, p. 2779. DOI: 10.3390/rs12172779.

21. Barriere V., Claverie M., Schneider M., Lemoine G., d’Andrimont R. Boosting crop classification by hierarchically fusing satellite, rotational, and contextual data. Remote Sensing of Environment, 2024, vol. 305, p. 114110 DOI: 10.1016/j.rse.2024.114110.

22. Karmakar P., Teng S.W., Murshed M., Pang S., Li Y., Lin H. Crop monitoring by multimodal remote sensing: A review. Remote Sensing Applications: Society and Environment, 2024, vol. 33, p. 101093 DOI: 10.1016/j.rsase.2023.101093.