Object Detection in Optical Remote Sensing Images: A Survey and A New Benchmark

By Ke Li, Gang Wan, Gong Cheng, Liqiu Meng, Junwei Han. Isprs Journal of Photogrammetry and Remote Sensing, 2019

2,173 citations recorded by the source corpus.

Substantial efforts have been devoted more recently to presenting various methods for object detection in optical remote sensing images. However, the current survey of datasets and deep learning based methods for object detection in optical remote sensing images is not adequate. Moreover, most of the existing datasets have some shortcomings, for example, the numbers of images and object categories are small scale, and the image diversity and variations are insufficient. These limitations greatly affect the development of deep learning based object detection methods. In the paper, we provide a comprehensive review of the recent deep learning based object detection progress in both the computer vision and earth observation communities. Then, we propose a large-scale, publicly available benchmark for object DetectIon in Optical Remote sensing images, which we name as DIOR. The dataset contains 23463 images and 190288 instances, covering 20 object classes. The proposed DIOR dataset 1) is large-scale on the object categories, on the object instance number, and on the total image number; 2) has a large range of object size variations, not only in terms of spatial resolutions, but also in the aspect of inter- and intra-class size variability across objects; 3) holds big variations as the images are obtained with different imaging conditions, weathers, seasons, and image quality; and 4) has high inter-class similarity and intra-class diversity. The proposed benchmark can help the researchers to develop and validate their data-driven methods. Finally, we evaluate several state-of-the-art approaches on our DIOR dataset to establish a baseline for future research.

Subfields: Computer Vision and Pattern Recognition, Media Technology, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/2dd479af74395f5731db202934c9cb02fae6e97b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

RSPrompter: Learning to Prompt for Remote Sensing Instance Segmentation Based on Visual Foundation Model (2023, background): “The applications of this technology span a multitude of domains, including, but not limited to, national defense, land surveying, disaster monitoring, and traffic planning [7], [10], [11], [12], [13], [14], [15], [16].”
SkySense: A Multi-Modal Remote Sensing Foundation Model Towards Universal Interpretation for Earth Observation Imagery (2023, methodology): “These tasks encompass scene classification using the AID dataset [80], object detection employing the DIOR dataset [40], semantic segmentation utilizing the Dyna.”
TTST: A Top-k Token Selective Transformer for Remote Sensing Image Super-Resolution (2024, neutral): “0 [64], DIOR [65] and NWPU-RESISC45 [66].”
RingMo: A Remote Sensing Foundation Model With Masked Image Modeling (2023, methodology): “1) Statistical Information: Some of the images in the dataset are selected from public datasets in the RS ﬁeld, such as [4], [5], [37], [40], [64], [65], and the other images are obtained from China’s Gaofen-2 satellite.”
Support Vector Machine Versus Random Forest for Remote Sensing Image Classification: A Meta-Analysis and Systematic Review (2020, background): “In addition, the availability of high-quality data, the temporal frequency, and comprehensive coverage make them advantageous for several agro-environmental applications compared to the traditional data collections approaches [6]–[8].”
Oriented R-CNN for Object Detection (2021, background): “In the past decade, remarkable progresses [23, 33, 2, 42, 31, 24, 45, 13, 20, 29] have been made in the field of object detection.”
Small-Object Detection in Remote Sensing Images with End-to-End Edge-Enhanced GAN and Object Detector Network (2020, result): “[58] showed review and comparison of different methods.”
Tiny Object Detection in Aerial Images (2021, background): “object detection benchmarks like DOTA [1] and DIOR [3] contain objects of various scales.”
FALSE: False Negative Samples Aware Contrastive Learning for Semantic Segmentation of High-Resolution Remote Sensing Image (2022, neutral): “DEEP neural networks (DNNs) trained in a supervised learning manner have made remarkable progress in remote sensing image (RSI) scene classification [1], target detection [2], and semantic segmentation [3], [4].”
Advancing Plain Vision Transformer Toward Remote Sensing Foundation Model (2022, methodology): “0 [43] and DIOR-R [44] for evaluation.”
Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges (2021, methodology): “However, the instances in xView and DIOR are both annotated by HBBs, which are ImageNet (2014 train) 200 456,567 82,820 478,807 1.05 DOTA-v1.”
Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, methodology): “Instance-Level Annotation: The general principle to annotate SODA resembles that of general detection benchmarks [6], [20], [30], [31], [32], and the only difference lies in the ignore regions.”
Oriented RepPoints for Aerial Object Detection (2021, methodology): “DIOR-R [2] gives the oriented bounding box annotations based on the DIOR dataset [16] for the oriented detection task.”
Incremental Detection of Remote Sensing Objects With Feature Pyramid and Knowledge Distillation (2020, methodology): “Therefore, several data sets have been developed in recent years, such as NWPU VHR-10 [10], DOTA [11], and DIOR [12].”
A Deeply Supervised Attention Metric-Based Network and an Open Aerial Image Dataset for Remote Sensing Change Detection (2021, background): “Inrecent years, the rapid rise of big data and the popu-larization of high computational power have promoted myriad developments in deep learning (DL), which has made remarkable achievements in many ﬁelds [14], [15], including remote sensing image interpretation [16]–[19].”
RemoteCLIP: A Vision Language Foundation Model for Remote Sensing (2023, methodology): “We combined six remote sensing datasets with object detection annotation, including DOTA [75], DIOR [76], HRRSD [77], RSOD [78], LEVIR [79], and HRSC [80].”
FAIR1M: A Benchmark Dataset for Fine-grained Object Recognition in High-Resolution Remote Sensing Imagery (2021, methodology): “As a result, most of existing object detection datasets in remote sensing are annotated with horizontal bounding boxes, such as UCAS-AOD [67], NWPU VHR-10 [7], DIOR [29], and SIMD [21].”
PANDA: Adapting Pretrained Features for Anomaly Detection and Segmentation (2020, methodology): “3 on FMNIST, DIOR, WBC, MVTec suggest that out-of-domain pretrained features are better at anomaly detection than indomain self-supervised features.”
Poly Kernel Inception Network for Remote Sensing Detection (2024, methodology): “PKINet-S is pretrained on ImageNet-1K [9] for 300 epochs which is consistent with previous methods [21, 32, 46] and built within the framework of Oriented RCNN [65]. mAP (07/12): VOC 2007 [13]/2012 [14] metrics.”