An Improved Aggregated-Mosaic Method for the Sparse Object Detection of Remote Sensing Imagery

Boya Zhao,Bing Zhang,Xinran Guan,Lianru Gao,Yuanfeng Wu

doi:10.3390/rs13132602

Abstract

Object detection based on remote sensing imagery has become increasingly popular over the past few years. Unlike natural images taken by humans or surveillance cameras, the scale of remote sensing images is large, which requires the training and inference procedure to be on a cutting image. However, objects appearing in remote sensing imagery are often sparsely distributed and the labels for each class are imbalanced. This results in unstable training and inference. In this paper, we analyze the training characteristics of the remote sensing images and propose the fusion of the aggregated-mosaic training method, with the assigned-stitch augmentation and auto-target-duplication. In particular, based on the ground truth and mosaic image size, the assigned-stitch augmentation enhances each training sample with an appropriate account of objects, facilitating the smooth training procedure. Hard to detect objects, or those in classes with rare samples, are randomly selected and duplicated by the auto-target-duplication, which solves the sample imbalance or classes with insufficient results. Thus, the training process is able to focus on weak classes. We employ VEDAI and NWPU VHR-10, remote sensing datasets with sparse objects, to verify the proposed method. The YOLOv5 adopts the Mosaic as the augmentation method and is one of state-of-the-art detectors, so we choose Mosaic (YOLOv5) as the baseline. Results demonstrate that our method outperforms Mosaic (YOLOv5) by 2.72% and 5.44% on 512 × 512 and 1024 × 1024 resolution imagery, respectively. Moreover, the proposed method outperforms Mosaic (YOLOv5) by 5.48% under the NWPU VHR-10 dataset.

Highlights

Object detection is a key component of remote sensing research [1,2,3] and plays a crucial role in the fields of urban planning, environment monitoring, and borderland patrol [4,5,6]
Input:Indetection net(), training samples ground for thistoepoch, resultsthe of validation dataset from last trainingafter model with poor results, validation dataset is evaluated theresult
We evaluate Aggregated-Mosaic for its capability to improve the object detection performance compared to other current methods

Summary

Introduction

Object detection is a key component of remote sensing research [1,2,3] and plays a crucial role in the fields of urban planning, environment monitoring, and borderland patrol [4,5,6]. Studies on object detection-based methods have made great progresses in recent years [7,8], for deep learning applications [9,10]. The dropout randomly deletes neural nodes or erases some regions. While the regional dropout method has improved the generalization of object detection tasks, the training of convolutional neural networks is a data-driven process, and current object detection methods are limited by a lack of training data. In the field of object detection, basic augmentation methods are typically combined with cropping, rotation, translation, and color jittering, and semantic augmentation meth-

Objectives

Methods

Results

Conclusion