Forestry Services

Drone Forest Survey

 We use drones to capture high-resolution imagery, then our own software identifies invasive plants and/or individual trees of the target stand within each forest subcompartment. Based on the analysis, we draw up an intervention plan so that silvicultural work can be targeted and carried out efficiently. |

Home Services Drone Forest Survey

Automated Detection and Control of Invasive Plants with Drone Surveys

The spread of invasive plants is one of the biggest challenges in forest management and nature conservation today. Tree of heaven, false indigo, box elder, common milkweed, cleavers and dewberry can take over large areas in a short time, pushing out native vegetation and making successful reforestation much harder.

Effective control starts with an accurate survey. On large sites that are hard to get through, though, tracking down invasive hotspots on foot takes a huge amount of time and labour, and checking the whole area is often simply not feasible.

That's why we developed our own drone survey and data processing system. Working from high-resolution RGB (visible spectrum) orthophotos, it automatically recognises and locates invasive plant species as well as individual trees of the target stand. After the survey we draw up an intervention plan and, if you wish, start controlling the invasive tree species.

Request a Quote
How Does the Method Work?

How Does the Method Work?

The system is based on analysing leaf morphology. Our own AI-based software examines high-resolution drone imagery for leaf arrangement, leaf shape, crown structure and other species-specific morphological features.

We build a separate database for each plant species, using data from our own field surveys. So the system doesn't do generic plant recognition – it learns the unique morphological traits of each individual species.

What makes our method effective is that the whole process – from field data collection and building the databases to the recognition software – is developed in-house. Each part builds on the others, which makes the entire workflow simpler, faster and more reliable.

Traditional vs Drone Forest Survey

Traditional vs Drone Forest Survey

Traditional field surveys are still essential for many forestry tasks, but on large or hard-to-access sites they can take a lot of time and labour. A drone-based forest health survey, on the other hand, offers the following advantages:

  • Faster coverage of large areas: data can be collected from larger, continuous forest areas in a short time.
  • Less fieldwork on foot: much of the area can be surveyed from the air, so there's no need to walk every part of it.
  • Access to hard-to-reach areas: difficult terrain is much less of an obstacle to data collection.
  • Consistent data: the entire survey area is captured with the same method and at high resolution, ready for analysis.
  • Automatic species recognition: with a suitable training database, specific invasive or target species can be recognised automatically.
  • Accurate mapping: GPS coordinates can be assigned to every identified plant and infested patch.
  • Easier change monitoring: repeat surveys using the same method show how the stand changes over time.
  • Support for treatment plans: the results can be used directly to produce plant-level treatment, thinning and spot-tending maps.
Who Is a Drone Forest Survey For?

Who Is a Drone Forest Survey For?

A drone-based forest health survey is especially useful if you want fast information on larger forest areas, collected with a consistent method and backed by accurate spatial data. The technology really pays off when the goal is to detect invasive plants, identify target-stand trees or plan later silvicultural work.

The service is used, among others, by:    

  • Forest managers: to detect invasive species, set treatment priorities, and plan tending and thinning work.  
  • Forest districts: for fast, consistent surveys of larger areas, documenting stand condition and preparing targeted interventions. 
  • Nature conservation organisations: to map the spread of invasive plants, track how it changes and support conservation management.  
  • Local councils: to assess the condition of the woodland and semi-natural areas they own or manage, and plan the work needed.  
  •  Research projects: to study the spatial distribution of plant species, carry out repeat surveys and collect high-resolution, georeferenced data.

        
Each survey is tailored to the site and the purpose of the study, so data collection and analysis focus only on the information you actually need to make decisions.

Built for Real Forestry Work

Built for Real Forestry Work

Our goal was a working system that performs reliably in everyday forestry surveys. We check our measurements continuously and use the feedback to keep improving the system.

Large-area drone surveys rarely come with perfect conditions. Shaded parts of the stand, changing cloud cover, canopy gaps, stands of different ages and the wind all affect image quality.

That's why we put at least as much effort into developing our data capture methodology as into the AI itself. After several years of development, we now have a robust survey method that allows surveys to be carried out even on 6–7-hour continuous working days, under changing light and in moderately windy weather.

We have also developed dedicated image processing methods to handle shaded parts of the stand, so the system isn't limited to evenly lit conditions.

Proven Accuracy

We first developed the method to detect tree of heaven (Ailanthus altissima). In real forestry surveys, we detect seed-bearing-sized tree of heaven with

97–100%

accuracy.

 

Contact
Other Applications

Other Applications

Setting Treatment Priorities

Not every data capture serves the same purpose. Imagery taken at the right phenological stage lets the system do more than pinpoint the plants of a given species – it can also analyse other traits that feed directly into forest management decisions.

For tree of heaven, for example, the system can pick out the seed-bearing trees. As these trees play the biggest role in creating new hotspots as the species spreads, they can be the top priority for treatment.

This makes it possible to:

  • separate the seed-bearing trees,
  • set treatment priorities.

Detecting Other Invasive Species

We have since applied what we learned while developing the system to other invasive plants. The system currently recognises the following automatically:

  • tree of heaven
  • black locust
  • box elder
  • common milkweed
  • wild blackberry

Mapping Target-Stand Trees

The method isn't limited to invasive plants, though. With a suitable species-specific training database, it can also locate individual trees of the target stand. This means a map of both invasive and target species can be produced from a single high-resolution orthophoto.

Survey Output

Survey Output

The system doesn't just recognise plants – it assigns a precise GPS coordinate to every identified plant or infested patch, which can be used directly in forestry operations.

Silviculture has always been our core business, so we've been able to test the applications listed below in practice, out in the field.

The data we deliver is used to produce, among other things:

  • plant-level treatment maps,
  • spot-tending plans,
  • thinning plans,
  • stand surveys.

Field crews can navigate straight to the area to be treated on a mobile phone or tablet, so no local knowledge of the site is needed.

Request a Quote
Continuous Development

Continuous Development

We still develop our system based on the results of real forestry surveys. Our aim isn't to recognise a single species, but to build a general methodology that, with the right training databases, can be adapted quickly to further invasive or native tree species.

Request a Quote

Our Other Services

Afforestation and Reforestation

We have been carrying out machine-based reforestation and afforestation for 12 years. Each year we regenerate or plant 150–200 hectares of forest with this technology. |

Learn more

Forest Tending

 Tending carried out at the right time and in the right way is one of the key conditions for successful reforestation. Its aim is to help the target stand develop and to keep competing and invasive plants in check. |

Learn more

Deer Fence Construction and Game Netting Installation

Protecting seedlings properly is one of the basics of successful regeneration. Large game can browse, trample and break young trees, causing serious damage to the stand. That’s why building a deer fence is often the most effective solution.

Learn more

Hawknest

HawkNest is a chemical mixing base developed specifically for drone spraying. We designed it around the typical needs of

agriculture, forestry and vineyards, and took the particular requirements of the Central and Eastern European market into account.

 

Learn more

Tree Planting Machine

 

In forestry planting, good root establishment is essential – especially on difficult terrain. We designed this machine specifically for the job: the right planting depth, special soil compaction and a robust build deliver reliable results over the long term.

 

Learn more

Sawdust Spreader

In tree nurseries, the right mulch layer is key to healthy young plants. Our sawdust spreader is built for exactly this job: it spreads sawdust or other organic mulch evenly and efficiently across the seedbeds.

Learn more
Contact

Request a quote from our forestry experts

Phone number: +36 30 712 0779 Email: info@mainforest.hu

Address: HU - 7960 Sellye, Carolina street 9.

Write to us: