Project

The forest is an important resource of both Latvia and other European countries. It provides both food and timber resource, plays a key role in carbon storage and climate change mitigation, as well as contributes to soil stabilisation and air purification. Forest planting is the most commonly used reforestation method - it is applicable to almost all types of forest growth conditions. It is important to carry out agrotechnical care in the first two to three years of an established coppice or forest plantation, as well as to identify dead trees early and replace them in time.
Agrotechnical care ensures the reduction of the gross coverage (shading) in the established coppice or forest plantation. It has a negative impact on tree saplings by delaying their growth by several years, contributing to their death, affecting the formation of crooked trunks, thus, promoting tree diseases, as well as increasing the population of various rodents that eat trees in the winter season. In order to prevent it, agrotechnical care must be carried out. Poorly executed agrotechnical management over a long period of time has a negative impact on the development of tree saplings and their annual wood growth. In Latvia, this control is carried out through human labour. Control is only partially automated, and the lack of human resources will have an increasing impact on the quality of control with each passing year, given that the area of young forest plantations continues to expand.
The aim of the project is to develop an automated unmanned aerial vehicle technology for monitoring agroforestry maintenance works of tree saplings. The new technology will allow foresters and potentially employees of the State Forest Service to carry out automated control of new forestry plantation maintenance work. The technology will also enable to determine the density of tree saplings and their extinction.
The project will develop automated unmanned aerial vehicle technology for monitoring coppicing works. This will allow to control the agrotechnical management in the established coppices or forest plantations using unmanned aerial vehicles and artificial intelligence methods. The UAV (commercially available) will carry out automated inspections of the forest area, where measurements from various sensors will be recorded and processed using artificial intelligence and computer vision techniques (project contribution).
The project idea is based on commercially available unmanned aerial vehicle technologies, open-source artificial intelligence technologies and computing capabilities, which, when integrated, will create a state-of-the-art solution to help solve challenges in the forestry sector. The proposed technology has significant potential for being used in future as well, given the decreasing availability of labour and the simultaneous rapid increase of established coppices and forest plantations. The DJI RTK UAV series offers a wide range of sensors available, as well as to control and program them as necessary. It will enable to map forest areas with help of open source and DJI's software, including calculation of terrain fluctuations, while the resolution of the visual sensors used will make fine details of tree saplings to appear visible.
Open source AI software libraries (e.g. PyTorch, TensorFlow, OpenCV with NVIDIA CUDA parallel computing technology) allow to create and further use the computer vision and machine learning data models to recognise objects and classify their properties in images. The market research showed that such technologies are not available within the European Union.

• Rīgas meži, SIA
• Latvijas Valsts mežzinātnes institūts "Silava"
• Mežsaimniecības pakalpojumu kooperatīvā sabiedrība "Mežsaimnieks", Kooperatīvā sabiedrība
• Sodra Latvia, SIA
• MANA MEŽS, SIA