The article provides a detailed review of the problem of deforestation, which in recent years has become uncontrolled. The main causes of forest damage are analyzed, among which the most well-known are climate change, diseases and pests. The losses of forestry as a result of tree diseases, which are large-scale and widespread in other countries, are given. The solution of these problems is possible under the condition of high-quality monitoring with the involvement of automated remote sensing tools and modern methods of image analysis, including artificial intelligence approaches such as neural networks and deep learning.
The article proposes an approach to automatic localization and recognition of trees affected by drought, which is of great practical importance for environmental monitoring and forestry. A fully connected convolutional model of deep learning using the tensorflow and keras libraries has been developed for localization and recognition. This model consists of a detector network and a separate classifier network. To train and test the proposed network based on images obtained by remote sensing, a training database containing 8500 images was created. A comparison of the proposed model with the existing methods is based on such characteristics as accuracy and speed. The accuracy and speed of the proposed recognition system were evaluated on a validation sample of images, consisting of 1700 images. The model has been optimized for practical use with CPU and GPU due to pseudo quantization during training. This helps to distribute the values of the weights in the learning process and bring their appearance closer to a uniform distribution law, which in turn allows more efficient application of quantization to the original model. The average operating time of the algorithm is also determined. In the Visual C++ environment, based on the proposed model, an expert program has been created that allows to perform the ecological monitoring and analysis of dry forests in the field in real time. Libraries such as OpenCV and Direct were used in software development, and the code supports object-oriented programming standards. The results of the work and the developed software can be used in remote monitoring and classification systems for environmental monitoring and in applied tasks of forestry.
- Linnakoski, R., Kasanen, R., Dounavi, A., & Forbes, M. (2019). Forest Health Under Climate Change: Effects on Tree Resilience, and Pest and Pathogen Dynamics. Frontiers in plant science, 3, 83–98. https://doi.org/10.3389/fpls.2019.01157
- Dukes, J., Pontius, J., Orwig, D., Garnas, J., Rodgers, V., Brazee, N., & Cooke, B. (2009) Responses of insect pests, pathogens, and invasive plant species to climate change in the forests of northeastern North America: What can we predict? Canadian Journal of Forest Research, 39(2), 231–248. https://doi.org/10.1139/X08-171
- Sturrock, R., Frankel, S., Brown, A., Hennon, P., Kliejunas, J., Lewis, K., Worrall, J., & Woods, A. (2011). Climate change and forest diseases. Plant Pathology, 60(1), 133–149. https://doi.org/10.1111/j.1365-3059.2010.02406.x
- Lechner, A., Foody, G., & Boyd, D. (2020). Applications in Remote Sensing to Forest Ecology and Management. One Earth, 2(5), 405–412. https://doi.org/10.1016/j.oneear.2020.05.001
- Halder1, M., Sarkar, A., & Bahar, H. (2019). Plant disease detection by image processing: A Literature review. Journal of Food Science & Technology 3(6), 534–538. https://doi.org/10.25177/JFST.3.6.6
- Arya, M., Anjali, K., & Unni, D. (2018). Detection of unhealthy plant leaves using image processing and genetic algorithm with Arduino. International Conference on Power, Signals, Control and Computation, 218–236. https://doi.org/10.1109/EPSCICON.2018.8379584
- Mehra, T., Kumar, V., & Gupta, P. (2016). Maturity and disease detection in tomato using computer vision. Fourth International Conference on Parallel, Distributed and Grid Computing, 68–72. https://doi.org/10.1109/PDGC.2016.7913228
- Rusyn, B., Lutsyk, O., Kosarevych, R., & Korniy, V. (2018). Segmentation of atmospheric clouds images obtained by remote sensing. 14th International Conference on Advanced Trends in Radioelectronics, Telecommunications and Computer Engineering, TCSET 2018, 213–216. https://doi.org/10.1109/TCSET.2018.8336189
- Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., & Chen, L. (2019). Inverted Residuals and Linear Bottlenecks: Mobile Networks for ClassiHcation, Detection and Segmentation. The IEEE Conference on Computer Vision and Pattern Recognition, Jan. 13, 4510–4520.
- Chng, C., & Chan, C. (2017). A comprehensive dataset for scene text detection and recognition. In Document Analysis and Recognition. 14th IAPR International Conference, 1, 935–942. https://doi.org/10.1109/ICDAR.2017.157