# Global Identification of Lunar Dark Mantle Deposits

## TL;DR

A new deep learning method identifies lunar dark mantle deposits, revealing their global distribution and composition.

## Contribution

A YOLOv8-based model with a multi-scale feature extraction module is introduced for identifying lunar dark mantle deposits.

## Key findings

- The DM-YOLO model achieved 83.9% precision, 83.2% recall, and 84.2% mAP@0.5 in identifying DMDs.
- Newly identified DMDs account for 9.2% of the global population, with 70.2% located in lunar highlands.
- The model's predictions showed strong consistency with known DMDs in chemical and mineralogical characteristics.

## Abstract

What are the main findings?
A YOLOv8-based deep learning method is proposed for the global identification of lunar dark mantle deposits (DMDs).Fifteen newly identified DMD candidates were validated using FeO abundance data and M3 spectral absorption features, demonstrating the reliability of the pre-dictions.

A YOLOv8-based deep learning method is proposed for the global identification of lunar dark mantle deposits (DMDs).

Fifteen newly identified DMD candidates were validated using FeO abundance data and M3 spectral absorption features, demonstrating the reliability of the pre-dictions.

What are the implications of the main findings?
Newly identified DMDs account for approximately 9.2% of the global DMD population.The updated global distribution shows that 29.8% of DMDs occur in mare regions, while 70.2% are located in the lunar highlands.

Newly identified DMDs account for approximately 9.2% of the global DMD population.

The updated global distribution shows that 29.8% of DMDs occur in mare regions, while 70.2% are located in the lunar highlands.

Lunar dark mantle deposits (DMDs), formed by explosive volcanic activity on the Moon, are typically composed of glass- and iron-rich pyroclastic materials, with slight variations in color, crystallinity, and TiO2 concentration by region. This paper proposes a method for identifying DMDs using the YOLOv8 deep learning model, enhanced by the introduction of a multi-scale feature extraction (MSFE) module with an attention mechanism, which improves the model’s ability to detect targets at different scales. First, a DMD dataset was constructed using Lunar Reconnaissance Orbiter (LRO) data, with manual annotations of DMD regions and lunar image slicing to optimize computational efficiency. The YOLOv8 architecture, with the incorporated MSFE module, was then used to improve model accuracy in complex terrain. The experimental results showed that the improved DM-YOLO model achieved a precision (P) of 83.9%, a recall (R) of 83.2%, and a mean average precision (mAP@0.5) of 84.2%, representing increases of 15.2%, 14.4%, and 14.0%, respectively, over those obtained with the original YOLOv8 model. The predicted results were preliminarily verified using FeO abundance data and further confirmed by analysis of M3 spectral absorption features, showing strong consistency with known DMDs in terms of both chemical composition and mineralogical characteristics. Observations showed that DMDs were located primarily in the low- and mid-latitude regions of the Moon, with most deposits found in the lunar highlands. The findings suggest that the DM-YOLO model has significant potential for providing technical support for lunar exploration and resource development, particularly for identifying small-scale features that are difficult to annotate.