Palm Tree Detection
Introduction
Palm trees are important economic crops in many tropical countries. The number of palm trees in a plantation area is important information for predicting the yield of palm oil, monitoring the growing situation of palm trees and maximizing their productivity, etc.
Traditionally, the counting of trees was done using physical survey or visual interpretation of imagery. An alternative technique is deep learning which can reduce time, resources and costs involved in such tasks. This technique can also be highly accurate.
Model description
- Input: This model is expected to perform with aerial imagery (5 - 15 cm resolution)
- Applicable geographies: This model is expected to work all over the world.
- Architecture: This model uses the FasterRCNN model architecture implemented in ArcGIS API for Python.
- Accuracy Metrics: This model has an average precision score of 0.75.
Map 1: Seluma, Indonesia
Table 1: Accuracy Matrices
Map 2: Kolovai, Tongatapu
Table 2: Accuracy Matrices
Map 3: Bengkulu, Indonesia
Table 3: Accuracy Matrices
Map 4: Near Luwu, Indonesia
Table 4: Accuracy Matrices
Map 5: Fatumu, Tongatapu
Table 5: Accuracy Matrices
Map 6: Tongatapu
Table 6: Accuracy Matrices
License requirements
The model can be downloaded by anyone who has an ArcGIS Online subscription. The following licenses are required to consume or run the model.
- ArcGIS Desktop: ArcGIS Image Analyst for ArcGIS Pro. The model cannot be used in ArcMap.
- ArcGIS Enterprise: ArcGIS Enterprise Map Viewer and ArcGIS Image Server with raster analytics configured.
- ArcGIS Online: ArcGIS Online Map Viewer using ArcGIS Online imagery.
Using the model
The model can be used with the Detect Objects Using Deep Learning tool.
