Object Recognition in Aerial Images Using Convolutional Neural Networks

Radović, Matija; Adarkwa, Offei; Wang, Qiaosong

doi:10.3390/jimaging3020021

Cited by 165 publications

(85 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Their network consists of 31,500 image chips related to 45 land classes and they reported an obvious improvement in scene classification by implementing this dataset on pre-trained networks [22]. Radovic et al worked on the detection of aircraft from unmanned aerial vehicle (UAV) imageries with YOLO and achieved a 99.6% precision rate [39]. Nie et al used SSD to detect the various sizes of ships inshore and offshore areas using a transfer-learned SSD model with 87.9% average precision and outperformed the Faster R-CNN model, which provided an 81.2% average precision [40].…”

Section: Introductionmentioning

confidence: 99%

Comparative Research on Deep Learning Approaches for Airplane Detection from Very High-Resolution Satellite Images

2020

View full text Add to dashboard Cite

Object detection from satellite images has been a challenging problem for many years. With the development of effective deep learning algorithms and advancement in hardware systems, higher accuracies have been achieved in the detection of various objects from very high-resolution (VHR) satellite images. This article provides a comparative evaluation of the state-of-the-art convolutional neural network (CNN)-based object detection models, which are Faster R-CNN, Single Shot Multi-box Detector (SSD), and You Look Only Once-v3 (YOLO-v3), to cope with the limited number of labeled data and to automatically detect airplanes in VHR satellite images. Data augmentation with rotation, rescaling, and cropping was applied on the test images to artificially increase the number of training data from satellite images. Moreover, a non-maximum suppression algorithm (NMS) was introduced at the end of the SSD and YOLO-v3 flows to get rid of the multiple detection occurrences near each detected object in the overlapping areas. The trained networks were applied to five independent VHR test images that cover airports and their surroundings to evaluate their performance objectively. Accuracy assessment results of the test regions proved that Faster R-CNN architecture provided the highest accuracy according to the F1 scores, average precision (AP) metrics, and visual inspection of the results. The YOLO-v3 ranked as second, with a slightly lower performance but providing a balanced trade-off between accuracy and speed. The SSD provided the lowest detection performance, but it was better in object localization. The results were also evaluated in terms of the object size and detection accuracy manner, which proved that large-and medium-sized airplanes were detected with higher accuracy.

show abstract

Section: Introductionmentioning

confidence: 99%

Comparative Research on Deep Learning Approaches for Airplane Detection from Very High-Resolution Satellite Images

2020

View full text Add to dashboard Cite

show abstract

“…One of the YOLO (You Only Look Once) studies proofed that this method, provides high accuracy n Object Recognition in Aerial Images Using Convolutional Neural Network, [6]. This study uses the YOLO method to detect aircraft from satellite imagery.…”

Section: Introductionmentioning

confidence: 93%

Pavement Distress Classification Using Deep Learning Method Based on Digital Image

Sulistyaningrum¹,

Oranova²,

Pramestya³

et al. 2020

Proceedings of the 2nd International Symposium on Transportation Studies in Developing Countries (ISTSDC 2019)

View full text Add to dashboard Cite

Maintaining the road regularly is a necessity, because the road is a vital infrastructure. One of automatic road maintenance steps is the detection of road distress type. Several methods have been used to detect and classify road distress automatically. This research determines the existence and classifies the deterioration of pavement using the Deep Learning method. The type of road distress detected are potholes, line-cracks, and non-line cracks. In this study, the deep learning method implemented is You Only Look Once (YOLO). The YOLO method uses Convolutional Neural Network (CNN) in its architecture and has given good results in object detection both on images and videos. YOLO has been tested in various datasets and given faster and accurate results. In this research, the pre-processing steps are cropping and resizing images then annotating the data. After that, the training is done by fine-tuning YOLO network process. The YOLO architecture uses 9-layer convolution and six layer maxpool. The testing results for datasets show that the highest accuracy is 99% and the highest average IoU is 75,1%. The run time of classification is 0.883 seconds per image.

show abstract

“…The authors show that their method performs better than the SSD model. In [49], the authors show that finding a suitable parameter setting can boost the object detection performance of convolutional neural networks on remote sensing imagery. They use YOLO [11] as object detector to optimize the parameters and infer the results.…”

Section: Object Detectionmentioning

confidence: 99%

Small-Object Detection in Remote Sensing Images with End-to-End Edge-Enhanced GAN and Object Detector Network

Rabbi

Ray

Schubert³

et al. 2020

Preprint

View full text Add to dashboard Cite

The detection performance of small objects in remote sensing images is not satisfactory compared to large objects, especially in low-resolution and noisy images. A generative adversarial network (GAN)-based model called enhanced super-resolution GAN (ESRGAN) shows remarkable image enhancement performance, but reconstructed images miss high-frequency edge information. Therefore, object detection performance degrades for small objects on recovered noisy and low-resolution remote sensing images. Inspired by the success of edge enhanced GAN (EEGAN) and ESRGAN, we apply a new edge-enhanced super-resolution GAN (EESRGAN) to improve the image quality of remote sensing images and use different detector networks in an end-to-end manner where detector loss is backpropagated into the EESRGAN to improve the detection performance. We propose an architecture with three components: ESRGAN, Edge Enhancement Network (EEN), and Detection network. We use residual-in-residual dense blocks (RRDB) for both the ESRGAN and EEN, and for the detector network, we use the faster region-based convolutional network (FRCNN) (two-stage detector) and single-shot multi-box detector (SSD) (one stage detector). Extensive experiments on a public (car overhead with context) and a self-assembled (oil and gas storage tank) satellite dataset show superior performance of our method compared to the standalone state-of-the-art object detectors.

show abstract

Object Recognition in Aerial Images Using Convolutional Neural Networks

Cited by 165 publications

References 11 publications

Comparative Research on Deep Learning Approaches for Airplane Detection from Very High-Resolution Satellite Images

Comparative Research on Deep Learning Approaches for Airplane Detection from Very High-Resolution Satellite Images

Pavement Distress Classification Using Deep Learning Method Based on Digital Image

Small-Object Detection in Remote Sensing Images with End-to-End Edge-Enhanced GAN and Object Detector Network

Contact Info

Product

Resources

About