Method for measuring tree diameters based on artificial intelligence technologies

In modern forestry, which emphasizes the sustainable use of resources, the key is the introduction of information technologies that provide accurate assessment of forest resources to ensure effective management, harvesting and processing. The research focuses on developing a tree trunk recognition method using modern technological solutions. The application of artificial intelligence (AI) technologies has significantly transformed approaches to measuring and analyzing physical objects, offering methods to automate the estimation of tree size and characteristics with increased accuracy and efficiency. The study emphasizes that optimal conditions for carrying out measurements using specialized technical means involve low tree density and the absence of abundant undergrowth and shrubs. The analysis shows that the integration of AI algorithms into data collection and analysis processes provides high measurement accuracy and reliability comparable to traditional manual methods, thus demonstrating its potential for practical application. The work identifies the problem of identifying individual trees when they are densely located, when the system may mistakenly perceive several trunks as a single object, which interferes with the accurate measurement of diameters. To further improve the accuracy and reliability of measurements, it is recommended to use unmanned aerial vehicles to collect visual data from different angles, develop and optimize AI algorithms, and conduct research on an expanded volume of data, which will help adapt AI technologies to the diverse conditions of forest ecosystems.

1. Ficko A. Bayesian Evaluation of Smartphone Applications for Forest Inventories in Small Forest Holdings. Forests, 2020, 11, р. 1148. DOI: 10.3390/f11111148.

2. Govyadin I.K. Innovative approaches to collecting data on the heights and diameters of trees in plantations. Prospects for the development of the forestry complex: collection of scientific papers of the international scientific and practical conference, Bryansk, December 18-19, 2023, рр. 179-182. EDN JWJABM. (In Russ.)

3. Govyadin I.K., Karimov B.M. Certificate of state registration of a computer program No. 2023686447 Russian Federation. Accounting and analysis system: No. 2023686474: application. 12/05/2023: publ. 12/06/2023; applicant Federal State Budgetary Educational Institution of Higher Education «Saint-Petersburg State Forest Technical University». EDN FJVAWI. (In Russ.)

4. Govyadin I.K., Karimov B.M., Sheremet V.A. Certificate of state registration of a computer program No. 2023612089 Russian Federation. Data visualization system: No. 2022668020: application. 09/30/2022: publ. 01/30/2023. EDN QVAQKV. (In Russ.)

5. Grondin V., Fortin J.-M., Pomerleau F., Giguere Ph. Tree Detection and Diameter Estimation Based on Deep Learning. 2022. DOI: 10.1093/forestry/cpac043.

6. Lebedev A.V. Inventory of tree plantations in urban areas using a smartphone. Forestry Journal, 2023, vol. 13, no. 3 (51), pp. 56-70. DOI: https://doi.org/10.34220/issn.2222-7962/2023.3/5. (In Russ.)

7. Portnov A.M., Ivanova N.V., Shashkov M.P. Experience of using the DeepForest neural network for detecting trees in a broad-leaved forest, Mathematical biology and bioinformatics: Reports of the IX International Conference, Pushchino, October 17-19, 2022, vol. 9, рр. 211-216. DOI: 10.17537/icmbb22.12. EDN: ODUVAY. (In Russ.)

8. Putra B., Indarto I., Marhaenanto B., Janna N. Yualianto, Soedibyo D. The use of computer vision to estimate tree diameter and circumference in homogeneous and production forests using a non-contact method. Forest Science and Technology, 2021, 17. DOI: 10.1080/21580103.2021.1873866.

9. Sandim A., Amaro M., Silva M., Cunha J., Morais S., Marques A., Ferreira A., Louzada J., Fonseca T. New Technologies for Expedited Forest Inventory Using Smartphone Applications. Forests, 2023, 14. DOI: 10.3390/f14081553.

10. Selyankin V.V. Computer vision. Image analysis and processing: textbook for universities. 2nd ed., erased. St. Petersburg: Lan, 2021. 152 p. (In Russ.)

11. Song C., Yang B., Zhang L. et al. A handheld device for measuring the diameter at breast height of individual trees using laser ranging and deep-learning based image recognition. Plant Methods, 2021, 17, 67. https://doi.org/10.1186/s13007-021-00748-z.

12. Tatsumi Sh., Yamaguchi K., Furuya N. Forest Scanner: A mobile application for measuring and mapping trees with LiDAR-equipped iPhone and iPad. Methods in Ecology and Evolution, 2022, 14. 10.1111/2041-210x.13900.

13. Voitov D.Yu., Vasilyev S.B., Kormilitsyn D.V. Development of technology for determining tree species using computer vision. Forest Bulletin, 2023, vol. 27, no. 1, pp. 60-66. DOI: 10.18698/2542-1468-2023-1-60-66. EDN: PXEBDF. (In Russ.)

14. Zhuk K.D., Ugryumov S.A. Selection of the optimal machine learning algorithm for the task of classifying tree trunk species. Digital technologies in the forest sector: materials of the IV All-Russian scientific and technical conference, St. Petersburg, October 19-20, 2023, pp. 32-34. EDN WAQLFH. (In Russ.)