Accurate Object Localization in Remote Sensing Images Based on Convolutional Neural Networks

By Yang Long, Yiping Gong, Zhifeng Xiao, Qing Liu. IEEE Transactions on Geoscience and Remote Sensing, 2017

783 citations recorded by the source corpus.

In this paper, we focus on tackling the problem of automatic accurate localization of detected objects in high-resolution remote sensing images. The two major problems for object localization in remote sensing images caused by the complex context information such images contain are achieving generalizability of the features used to describe objects and achieving accurate object locations. To address these challenges, we propose a new object localization framework, which can be divided into three processes: region proposal, classification, and accurate object localization process. First, a region proposal method is used to generate candidate regions with the aim of detecting all objects of interest within these images. Then, generic image features from a local image corresponding to each region proposal are extracted by a combination model of 2-D reduction convolutional neural networks (CNNs). Finally, to improve the location accuracy, we propose an unsupervised score-based bounding box regression (USB-BBR) algorithm, combined with a nonmaximum suppression algorithm to optimize the bounding boxes of regions that detected as objects. Experiments show that the dimension-reduction model performs better than the retrained and fine-tuned models and the detection precision of the combined CNN model is much higher than that of any single model. Also our proposed USB-BBR algorithm can more accurately locate objects within an image. Compared with traditional features extraction methods, such as elliptic Fourier transform-based histogram of oriented gradients and local binary pattern histogram Fourier, our proposed localization framework shows robustness when dealing with different complex backgrounds.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/cac077ce6700f9bb3cdc2fc5dd1dfe4aefb5f149

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

Citations in the corpus

Rotation-Insensitive and Context-Augmented Object Detection in Remote Sensing Images (2018, methodology): “Selective search has been broadly used as the region proposal method by many state-of-the-art geospatial object detectors, including [14], [15].”
DOTA: A Large-Scale Dataset for Object Detection in Aerial Images (2017, background): “Extensive studies have been devoted to object detection in aerial images [24, 15, 18, 3, 20, 39, 19, 32, 31, 22], drawing upon recent advances in Computer Vision and accounting for the high demands of Earth Vision applications.”
Towards Geospatial Foundation Models via Continual Pretraining (2023, neutral): “NAIP [31] 300,000 1m n/a RSD46-WHU [27] 116,893 0.”
PatternNet: A Benchmark Dataset for Performance Evaluation of Remote Sensing Image Retrieval (2017, background): “…success of deep learning, the remote sensing community has realized the potential of applying deep learning techniques for remote sensing tasks such as scene classification [34–38], object detection [39–41], semantic segmentation [42,43], image super-resolution [44] and image retrieval [45–49].”
Object Detection in Optical Remote Sensing Images: A Survey and A New Benchmark (2019, methodology): “(Long et al., 2017) presented an unsupervised score‐based bounding box regression (USB‐BBR) method based on R‐ CNN framework.”
CLRS: Continual Learning Benchmark for Remote Sensing Image Scene Classification (2020, background): “In the same way as most of the existing datasets images are collected, such as AID++ [23], RSD46-WHU [32,33], etc., CLRS images are mainly collected from Google Earth, Bing Map, Google Map, and Tianditu, which use different remote imaging sensors.”
A Survey of Deep Learning-Based Object Detection (2019, background): “[153] concentrate on automatically and accurately locating objects.”
Self-Supervised Remote Sensing Feature Learning: Learning Paradigms, Challenges, and Future Works (2022, background): “a) RSD46-WHU [62], [251], a large-scale scene classification dataset containing 117 000 samples of 46 categories.”
EarthGPT: A Universal Multimodal Large Language Model for Multisensor Image Comprehension in Remote Sensing Domain (2024, methodology): “The optical datasets include DIOR [61], DOTA [62], FAIR1M [63], HRRSD [64], NWPUVHR10 [65], RSOD [85], UCAS-AOD [86], and VisDrone [87].”
Learning a Rotation Invariant Detector with Rotatable Bounding Box (2017, methodology): “A CNN based pio-neering study upon vehicle detection in satellite images is presented in[2], followed by many approaches that focus on different types of geospatial object detection problems [14, 20, 21, 5, 26, 17].”
Learning RoI Transformer for Oriented Object Detection in Aerial Images (2019, methodology): “With more and more aerial images being available, object detection in aerial images has been a speciﬁc but active topic in computer vision [3, 29, 36, 6].”
Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges (2021, background): “some public datasets [17], [42], [43], [44], [45] have multiple categories, they have only limited number of samples, which are hardly efficient for training robust deep models.”
A Transdisciplinary Review of Deep Learning Research and Its Relevance for Water Resources Scientists (2017, methodology): “…an image for what they are, e.g., land use classes and crops types; Geng et al., 2015), object recognition (finding targets from a series of images; Wagner, 2016), object localization (Long et al., 2017), and terrain attribute extraction, for example, sea ice concentration (L. Wang et al., 2016).”