Machine Learning Stocking Assessment

Interpine has developed a complete solution for accurate tree counts and stocking assessments, using a machine learning (ML) model to detect and count trees in aerial imagery or LiDAR data. This model requires little effort to train and provides highly accurate results. 

We can undertake the drone survey, or you can collect the aerial imagery/LiDAR with your own drone, following our recommended specs below. Get in touch if you have any questions.

Post Harvest Waste Detection

We can also apply our machine learning detection tools to assess the amount of residual waste in a stand after harvesting.  This is a useful approach for detecting stems that may have been missed during harvesting, or for assessing the amount of slash near waterways.   

Map your harvest area following the specs below, or get us to do it.

Data Specifications:

RGB Cameras: 

The following UAV’s can be used: DJI P4 / DJI P4V2 / DJI P4 RTK / DJI Mavic 2
The following cameras can be used: DJI P1 / DJI X7 / DJI X5S

RGB Flight Specs: 

  • 80% front overlap
  • 80% side overlap
  • 120m height AGL (max)
  • Flight speed = 2 x GSD (e.g. speed 6m/s for 3cm GSD)
  • Exposure locked (not on auto)

LiDAR Sensors: 

The following units can be used: DJI Zenmuse L1, Yellowscan Voyager, Yellowscan Mapper+ LiDAR, Yellowscan Explorer, Nextcore Lumos XM120

LiDAR Flight Specs: 

  • 70% side overlap
  • 100m height AGL (max)
  • Flight speed = 10m/s (max)
  • Triple return
  • 120mhz
  • RGB colouring on if possible