Krasnodar, Krasnodar, Russian Federation
UDC 004.94
CSCSTI 36.01
The article presents the results of a study on the selection of an optimal neural network for identifying agricultural land plots, illustrated with vineyards, for updating the state agricultural land registry. The primary goal of the inventory is to detect any inconsistencies between the federal vineyard registry data and the vineyards’ current condition. For the experiments, the YOLOv5, YOLOv8, and Mask R-CNN neural networks were selected as the most commonly used for the purposes of recognizing objects in images. Neural networks offer advanced methods of image analysis that can be used for automated identification of vineyards on agricultural land. Their performance in the detection task was determined by calculating the Precision, Recall, and mAP metrics on a validation data set. The results of the comparison of the YOLOv5, YOLOv8, and Mask R-CNN models are presented in the table, which can be used to track their effectiveness. A comparison of the models showed that for a set of agricultural land plot images with vineyards, the YOLOv5 model may be preferable for tasks that require gradual improvement with an increasing number of epochs, while Mask R-CNN provides consistent high results even with a small number of epochs. YOLOv8 lags significantly in all metrics, especially in the early stages, and demonstrates the lowest overall performance.
agricultural land plots, land plot identification, vineyards, neural networks (NN/ANN), image detection, artificial intelligence, Mask R-CNN, YOLOv8, YOLOv5, metrics, mAP, Precision, Recall
1. Kuznecov P.N., Kotel'nikov D.Yu. Avtomatizirovannyy tehnologicheskiy kompleks monitoringa i diagnostiki vinogradnikov // Vestnik agrarnoy nauki Dona. 2021. T. 4. № 56. S. 16–23.
2. Studenkova N.A., Dobrotvorskaya N.I. Sovremennye problemy inventarizacii i kadastrovogo ucheta zemel' sel'skohozyaystvennogo naznacheniya // Sbornik materialov XVII Mezhdunarodnogo nauchnogo kongressa «Interekspo GEOSibir'»: v 8 t. Novosibirsk: SGUGiT, 2021. T. 3. № 2. S. 198–204. DOIhttps://doi.org/10.33764/2618-981X-2021-3-2-198-204.
3. Kuznecov P.N., Kotel'nikov D.Yu., Voronin D.Yu. Neyrosetevoe avtomatizirovannoe detektirovanie priznakov uhudsheniya sostoyaniya vinogradnyh nasazhdeniy // Problemy nauchnoy mysli. 2023. T. 2. № 8. S. 48–55.
4. Kuznecov P.N., Kotel'nikov D.Yu., Voronin D.Yu. Tehnologiya avtomatizirovannogo monitoringa sostoyaniya vinogradnika // Agrarnaya nauka. 2023. № 3. S. 109–116. DOIhttps://doi.org/10.32634/0869-8155-2023-368-3-109-116.
5. Wang T., Gan V.J. Automated joint 3D reconstruction and visual inspection for buildings using computer vision and transfer learning // Automation in Construction. 2023. Vol. 149. 104810. DOIhttps://doi.org/10.1016/j.autcon.2023.104810.
6. Gura D.A., Markovskiy I.G., Pshidatok S.K. Metodika monitoringa ob'ektov nedvizhimosti s pomosch'yu trehmernogo lazernogo skanirovaniya v specifike gorodskih zemel' // Geodeziya i kartografiya. 2021. T. 82. № 4. S. 45–53. DOIhttps://doi.org/10.22389/0016-7126-2021-970-4-45-53.
7. Dhanya V.G., Subeesh A., Kushwaha N.L., et al. Deep learning based computer vision approaches for smart agricultural applications // Artificial Intelligence in Agriculture. 2022. Vol. 6. P. 211–229. DOIhttps://doi.org/10.1016/j.aiia.2022.09.007.
8. D'yachenko R.A., Dovgal' V.V., Gura D.A. K voprosu sravneniya effektivnostineyronnyh setey YOLOv8 i U-Net v zadachah segmentacii territorial'nyh ob'ektov // Informacionnye tehnologii. 2024. T. 30. № 9. S. 480–485. DOIhttps://doi.org/10.17587/it.30.480-485.
9. Antyufeev V.V., Ryabov V.A. Opyt agroklimatologicheskogo obosnovaniya proektov plodovyh nasazhdeniy v Severnom Prichernomor'e v epohu global'nogo potepleniya // Izvestiya Orenburgskogo gosudarstvennogo agrarnogo universiteta. 2017. T. 4. № 66. S. 252–256.
10. Filonova M.A., Horoshko M.B. Primenenie sredstv mashinnogo zreniya v vinogradnyh hozyaystvah: podgotovka dannyh i obuchenie modeli // Upravlenie bol'shimi sistemami: sbornik nauchnyh trudov XIX Vserossiyskoy shkolykonferencii molodyh uchenyh (Voronezh, 5–8 sentyabrya 2023 g.). Voronezh: VGTU, 2023. S. 417–424.
11. Krizhevsky A., Sutskever I., Hinton G.E. ImageNet Classification with Deep Convolutional Neural Networks // Proceedings of the 25th International Conference on Neural Information Processing Systems (Lake Tahoe, Nevada, Dec. 2012). 2012. Vol. 1. P. 1097–1105.
12. Wang, C.Y., Bochkovskiy A., Liao H.Y.M. CSPNet: A new backbone that can enhance learning capability of CNN // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Seattle, WA, USA, 2020. P. 1571–1580. DOIhttps://doi.org/10.1109/CVPRW50498.2020.00203.
13. Liu S., Qi L., Qin H., et al. Path Aggregation Network for Instance Segmentation // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Salt Lake City, UT, USA, 2018. P. 8759–8768. DOIhttps://doi.org/10.1109/CVPR.2018.00913.
14. Ren S., He K., Girshick R., et al. Faster R-CNN: Towards Real-Time Object Detection withRegion Proposal Networks // Proceedings of the IEEE Transactions on Pattern Analysis and Machine Intelligence (June 1, 2017). 2017. Vol. 39. No. 6. P. 1137–1149. DOI:10.1109/ TPAMI.2016.2577031.
15. Girshick R. Fast R-CNN // Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV). Santiago, Chile, 2015. P. 1440–1448. DOIhttps://doi.org/10.1109/ICCV.2015.169.
16. He K., Gkioxari G., Dollár P., et al. Mask R-CNN // Proceedings of the IEEE InternationalConference on Computer Vision (ICCV). Venice, Italy, 2017. P. 2980–2988. DOIhttps://doi.org/10.1109/ICCV.2017.322.
17. Paletta Q., Terrén-Serrano G., Nie Y., et al. Advances in solar forecasting: Computer vision with deep learning // Advances in Applied Energy. 2023. Vol. 11. 100150. DOIhttps://doi.org/10.1016/j.adapen.2023.100150.
18. Talaat F.M., ZainEldin H. An improved fire detection approach based on YOLO-v8 for smart cities // Neural Comput & Applic. 2023. Vol. 35. P. 20939–20954.DOIhttps://doi.org/10.1007/s00521-023-08809-1.
19. Lipilin D.A., Evtushenko D.D. Ocenka kachestva gorodskoy sredy s primeneniem geoinformacionnyh sistem na primere mikrorayonov goroda Krasnodara // Geologiya i geofizika Yuga Rossii. 2022. T. 12. № 3. S. 195–210. DOIhttps://doi.org/10.46698/VNC.2022.72.93.013.
20. Pogorelov A.V., Laguta A.A., Netrebin P.B., et al. Analysis of the bottom topography of the reservoir due to sediment trapping (according to the Krasnodar reservoir, Russia) // Geography, Environment, Sustainability. 2023. Vol. 16. No. 3. P. 102–112.
21. Chengjun Chen, Chunlin Zhang, Jinlei Wang, et al. Semantic segmentation of mechanical assembly using selective kernel convolution UNet with fully connected conditional random field // Measurement. 2023. Vol. 209. 112499. DOIhttps://doi.org/10.1016/j.measurement.2023.112499.
