Skip to content
Deep Learning - Palm Tree Detection
In this project, I used a pretrained deep learning model in ArcGIS Pro to detect palm trees from high-resolution drone imagery of Kolovai on Tongatapu. Instead of training a model from scratch, I applied the Palm Tree Detection model from ArcGIS Living Atlas to create a feature layer of detected palm trees.
Figure 1. ArcGIS Pro is prepared here for the palm tree detection workflow, with the Kolovai imagery loaded and the Detect Objects Using Deep Learning geoprocessing tool open. The processing extent is set to the selected area where the pretrained model will be run.
Figure 2. The completed palm tree detection output is shown here with yellow outlined polygons over the imagery.
The result of the workflow was a palm tree detection layer shown as yellow outlined polygons over the imagery. After the model was run, I adjusted the symbology so the detections were easier to see and review on the map. The final result showed that most palm trees in the selected area were detected clearly, with only a small number of incorrect detections
Conclusion
Overall, this project showed how a pretrained model can quickly identify palm trees in imagery and create a clear detection layer. It was quite surprising how accurate this tool was, most of the palm trees in the image were detected well with only a small number of missed or incorrect detections. The finished output was a simple way to map the palm trees across the area without having to identify each one by hand.
Want to print your doc?
This is not the way.
This is not the way.
Try clicking the ··· in the right corner or using a keyboard shortcut (
CtrlP
) instead.