Science Geomatics Notes & Supplements

GIS tutorials, geospatial tools, palm plantation mapping, and custom software for the geomatics community.

50+
Tutorials
10+
Tools
9
Courses

Tuesday, 22 September 2026

Palm Counter Pro: Count 100,000 Oil Palm Trees in Minutes, Not Weeks

Counting oil palms is one of those jobs everyone hates.

If you've ever done it manually, you know: you open the orthomosaic in QGIS, create a point layer, and start clicking. One palm at a time. For a 100-hectare estate with ~25,000 palms, that's days of clicking  and your accuracy drops off a cliff by hour two. Hire someone to do it and you'll pay them to be bored and make mistakes.

I've spent the last few years doing palm census work for plantations here in Malaysia, and I got tired of this too. So I built the tool I always wanted.

Meet Palm Counter Pro

Palm Counter Pro is a desktop application that automatically detects and counts oil palm trees from drone (UAV) orthomosaic imagery. You give it an orthophoto, it gives you every palm as a georeferenced point — with crown delineation, per-palm statistics, and a full report. And because it was built for real plantation health monitoring, it goes one step further: an optional module predicts Ganoderma (basal stem rot) infection risk per palm from multispectral features.

No cloud subscription. No per-image fees. No internet needed. It runs on your own Windows PC, on your own data.


Real detection output: every palm crown automatically delineated (red circles) from UAV orthomosaic.

What It Does

  • Automatic detectionYOLO-based palm crown detection from RGB orthomosaic — even in mature, overlapping canopy. Confidence & IoU fully adjustable.
  • Every palm, georeferencedOutputs a GeoPackage (GPKG) point layer with real-world coordinates, ready for QGIS/ArcGIS.
  • Crown delineationGenerates crown circles and per-palm histograms for every detected palm.
  • Handles big filesTiled processing for large orthomosaics — no RAM choke, no size limits.
  • GPU acceleratedRuns on NVIDIA GPUs (CUDA) with batch-size control — a 100 ha estate in minutes, not days.
  • Runs 100% offlineYour plantation data never leaves your computer.

How It Works — 3 Steps



The application: 3-step pipeline — AI palm detection → crown delineation & feature extraction → optional disease prediction.

  1. Load your orthomosaicGeoTIFF from Agisoft Metashape, Pix4D, DJI Terra, WebODM — any of them.
  2. Press StartThe engine tiles the image, detects palm crowns, and filters false positives automatically.
  3. ExportPalm points CSV, statistics report, and annotated map — all generated for you.

Typical processing time: under 5 minutes for a 100-hectare estate on a normal office PC.

Why Not Just Count Manually?

I ran the comparison on a real estate:

MethodTimeCostAccuracy after hour 3
Manual clicking3–5 daysStaff time × daysDrops fast (fatigue)
Palm Counter Pro~5 minOne-time licenseConsistent across the whole estate

The tool doesn't get tired on palm number 18,000. It counts block 40 with the same consistency as block 1.

Who It's For

  • Estate managersCensus, replanting verification, density audits.
  • GIS vendors & survey companiesDeliver palm census as a service without hiring clickers.
  • Agronomists & researchersStand count trials, canopy studies.
  • SmallholdersKnow exactly how many palms you own.

Technical Specs

  • Input: GeoTIFF orthomosaic (RGB), any CRS, 2–10 cm/px recommended — from Agisoft Metashape, Pix4D, DJI Terra, WebODM
  • Detection engine: YOLO (.pt or .onnx model), adjustable confidence & IoU, global NMS, RGB band mapping
  • Output: GeoPackage (GPKG) palm points · crown circles · per-palm histograms · optional multispectral feature extraction
  • Bonus module: XGBoost-based Ganoderma (basal stem rot) infection prediction from multispectral features
  • Platform: Windows 10/11 · NVIDIA GPU recommended (CUDA) · 8 GB RAM minimum
  • No installation headaches: single .exe, no Python, no dependencies

Pricing

RM 499
one-time · single user · lifetime license · free updates for 1 year

For companies: multi-seat and enterprise licensing available. I'll send a proper quotation for your PO process.

Company purchase? Ask for a formal quotation — PO and bank transfer accepted.

 If it can handle Malaysian oil palm estates, it can handle yours.

Sunday, 26 November 2017

Lab Exercise 2: Animated Cartogram - Geovisualization (GLS688)





   Cartogram is a unique type of map because it combines statistical information with geographic location. Physical or topographical maps show relative area, distance, and terrain, but they do not provide any data about the inhabitants of a place. Cartograms on the other hand take some measurable variable: total population, age of inhabitants, electoral votes, GDP, etc., and then manipulate a place’s area to be sized accordingly. The produced cartogram can really look quite different from the maps of cities, states, countries, and the world that are more recognizable. It all depends on how a cartographer needs or wants to display the information. Cartograms come in all shapes and sizes, literally, and with the continuous advances in technology of Geographic Information System(GIS) software. This post presents a rough idea of how one might create a cartogram using QGIS.


Data Download :




Project 1 : Interactive Animation and Web Mapping of Sabah Earthquake from 1970 to 2016 - Geovisualization (GLS688)




Map layers may contain temporal data an attribute field containing the date and time that something happened. These temporal events can occur in the same place, or can occur in different places, over the course of time.  In the web map, temporal data can be visualized using a time slider. The time slider provides controls to explore temporal data interactively and is available at the bottom of any map that contains enabled temporal layers. Time Aware is a configurable app template that enables users to visualize time-enabled layers. It enhances the temporal capabilities of the web map, and can be configured to display time-enabled web maps in different ways.


Data Download




Procedure :

Sunday, 8 October 2017

Lab Exercise 1: Mapping Burglary Hotspot in Shah Alam - Geovisualization (GLS688)


  Heatmaps are one of the best visualization tools for dense point data. Heatmaps are utilized to effectively recognize discover groups where there is a high convergence of movement. They are likewise valuable for doing cluster analysis and hotspot examination.

    We will work with a dataset of burglary activites in Shah Alam and find crime hotspots in the county by performing HotSpot or Cluster analysis on dense point data. It will answer the highest burglary hotspot in that location and which police jurisdiction need to improvise and develop more of its branches.


Data Download :



Procedure :





Friday, 6 October 2017

Visualize Bird Migration Data in QGIS - Geovisualization (GLS688)



    A common use case of the QGIS TimeManager plugin is visualizing tracking data such as animal migration data. This post illustrates the steps necessary to create an animation from bird migration data.


Data Download :



Procedure :




Thursday, 5 October 2017

Lab 2 Creating Heatmaps using QGIS - Geovisualization (GLS688)



    Heatmaps are one of the best visualization tools for dense point data. Heatmaps are utilized to effectively recognize discover groups where there is a high convergence of movement. They are likewise valuable for doing cluster analysis and hotspot examination.

    We will work with a dataset of crime locations in Surrey, UK for the year 2011 and find crime hotspots in the county by performing HotSpot or Cluster analysis on dense point datadata.police.uk provides street-level crime, outcome, and stop and search data in simple CSV format. Download the data for Surrey Police and unzip the downloaded archive to extract the CSV file. 


Data Download :



Procedure : 





Thursday, 28 September 2017

Lab 1 Revisit John Snow's Map 1854 using QGIS - Geovisualization (GLS 688)


    John Snow (15 March 1813 – 16 June 1858) was an English physician and a leader in the adoption of anaesthesia and medical hygiene. He is considered one of the fathers of modern epidemiology, in part because of his work in tracing the source of a cholera outbreak in Soho, London, in 1854. His findings inspired fundamental changes in the water and waste systems of London, which led to similar changes in other cities, and a significant improvement in general public health around the world.

     This QGIS tutorial walks through a John Snow's famous cholera investigation in Soho with a 3D surface using Qgis2threejs plug-in, raster creation tools and Spatial Analyst. There are multiple sources for the similar data used in this presentation, including data compiled by Robin Wilson.

Download Data Link :



Procedure :



Final Result :