Detection and Geographic Localization of Natural Objects in the Wild: A Case Study on Palms

TL;DR

This paper introduces PRISM, a comprehensive pipeline for detecting and mapping palms in dense tropical forests using large aerial imagery datasets, with potential adaptation for other natural objects.

Contribution

The paper presents a new large UAV-derived dataset, evaluates advanced object detection methods, and refines geographic mapping techniques for natural object detection in complex forest environments.

Findings

PRISM effectively detects palms in dense forests.

Zero-shot SAM 2 improves segmentation accuracy.

Calibration enhances confidence score reliability.

Abstract

Palms are ecologically and economically indicators of tropical forest health, biodiversity, and human impact that support local economies and global forest product supply chains. While palm detection in plantations is well-studied, efforts to map naturally occurring palms in dense forests remain limited by overlapping crowns, uneven shading, and heterogeneous landscapes. We develop PRISM (Processing, Inference, Segmentation, and Mapping), a flexible pipeline for detecting and localizing palms in dense tropical forests using large orthomosaic images. Orthomosaics are created from thousands of aerial images and spanning several to hundreds of gigabytes. Our contributions are threefold. First, we construct a large UAV-derived orthomosaic dataset collected across 21 ecologically diverse sites in western Ecuador, annotated with 8,830 bounding boxes and 5,026 palm center points. Second, we evaluate multiple state-of-the-art object detectors based on efficiency and performance, integrating zero-shot SAM 2 as the segmentation backbone, and refining the results for precise geographic mapping. Third, we apply calibration methods to align confidence scores with IoU and explore saliency maps for feature explainability. Though optimized for palms, PRISM is adaptable for identifying other natural objects, such as eastern white pines. Future work will explore transfer learning for lower-resolution datasets (0.5 to 1m).