Vehicle and Vessel Detection on Satellite Imagery: A Comparative Study on Single-Shot Detectors

Ophoff, Tanguy; Puttemans, Steven; Kalogirou, Vasileios; Robin, Jean-Philippe; Goedemé, Toon

doi:10.3390/rs12071217

Cited by 22 publications

(13 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our framework, we aim for an object detector designed to perform well on aerial images. Although many object detectors specialized for aerial images have been proposed in the past (see, e.g., [4], [50]), we observe that standard object detectors with small modifications are suitable for our data. We decide to use the RetinaNet [7] for this work and modify it accordingly.…”

Section: Object Detectormentioning

confidence: 90%

Artificial and beneficial – Exploiting artificial images for aerial vehicle detection

Weber

Bongartz

Roscher

2021

ISPRS Journal of Photogrammetry and Remote Sensing

View full text Add to dashboard Cite

Object detection in aerial images is an important task in environmental, economic, and infrastructure-related tasks. One of the most prominent applications is the detection of vehicles, for which deep learning approaches are increasingly used. A major challenge in such approaches is the limited amount of data that arises, for example, when more specialized and rarer vehicles such as agricultural machinery or construction vehicles are to be detected. This lack of data contrasts with the enormous data hunger of deep learning methods in general and object recognition in particular. In this article, we address this issue in the context of the detection of road vehicles in aerial images. To overcome the lack of annotated data, we propose a generative approach that generates top-down images by overlaying artificial vehicles created from 2D CAD drawings on artificial or real backgrounds. Our experiments with a modified RetinaNet object detection network show that adding these images to small real-world datasets significantly improves detection performance. In cases of very limited or even no real-world images, we observe an improvement in average precision of up to 0.70 points. We address the remaining performance gap to real-world datasets by analyzing the effect of the image composition of background and objects and give insights into the importance of background.

show abstract

Section: Object Detectormentioning

confidence: 90%

Artificial and beneficial – Exploiting artificial images for aerial vehicle detection

Weber

Bongartz

Roscher

2021

ISPRS Journal of Photogrammetry and Remote Sensing

View full text Add to dashboard Cite

show abstract

“…When applying object detection on orthomosaics, the general approach is to cut the image into smaller patches, which are then individually processed by the deep learning pipeline [3]. Recently, several advancements have been made towards aerial object detection on orthomosaics [1,12,18]. Networks that perform less subsampling, or upsample their features with deconvolutions or feature pyramids tend to perform better, as the objects in aerial photography are usually smaller than their counterparts in natural images.…”

Section: Related Workmentioning

confidence: 99%

Improving Object Detection in VHR Aerial Orthomosaics

Ophoff

Beeck

Goedemé

2023

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

In this paper we investigate how to improve object detection on very high resolution orthomosaics. For this, we present a new detection model ResnetYolo, with a Resnet50 backbone and selectable detection heads. Furthermore, we propose two novel techniques to post-process the object detection results: a neighbour based patch NMS algorithm and an IoA based filtering technique. Finally, we fuse color and depth data in order to further increase the results of our deep learning model. We test these improvements on two distinct, challenging use cases: solar panel and swimming pool detection. The images are very high resolution color and elevation orthomosaics, taken from plane photography. Our final models reach an average precision of 78.5% and 44.4% respectively, outperforming the baseline models by over 15% AP.

show abstract

“…[79] Comparison among faster R-CNN, R-FCN, and SSD (Best model) [80] Optimized DL model considering feature extraction, object detection, and non-maximum suppression. [81] Small-Sized Vehicle Detection Network (AVDNet) (one-stage vehicle detection network) [82] Comparison among four object detection networks: D-YOLO (best model), YOLOV2, YOLOV3, and YOLT [83] Vehicle detection based on RetinaNet architecture [84] Model based on Alexnet network (classification) and Faster R-CNN (target detection) [85] Faster R-CNN with a improved feature-balanced pyramid network (FBPN) [86] Comparison among YOLOv3, YOLOv4 (best models), and Faster R-CNN [87] Super-resolution cyclic GAN with RFA and YOLO as the detection network (SRCGAN-RFA-YOLO) 1, 2 [88] Modified YOLOv3 and fcNN using 3D features in cascade. [16] Method using the lightweight feature extraction network with the Faster R-CNN [17] Orientation-Aware Vehicle Detection with an Anchor-Free Object Detection approach…”

Section: Papermentioning

confidence: 99%

Bounding Box-Free Instance Segmentation Using Semi-Supervised Iterative Learning for Vehicle Detection

Carvalho

Albuquerque

Santana

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Vehicle classification is a hot computer vision topic, with studies ranging from ground-view to top-view imagery. Topview images allow understanding city patterns, traffic management, among others. However, there are some difficulties for pixel-wise classification: (a) most vehicle classification studies use object detection methods, and most publicly available datasets are designed for this task, (b) creating instance segmentation datasets is laborious, and (c) traditional instance segmentation methods underperform on this task since the objects are small. Thus, the present research objectives are: (1) propose a novel semi-supervised iterative learning approach using GIS software, (2) propose a box-free instance segmentation approach, and (3) provide a city-scale vehicle dataset. The iterative learning procedure considered: (1) labeling a few vehicles from the entire scene, (2) choosing training samples near those areas, ( 3) training the DL model (U-net with Efficient-net-B7 backbone), (4) classifying the whole scene, (5) converting the predictions into shapefile, (6) correcting areas with wrong predictions, (7) including them in the training data, (8) repeating until results are satisfactory. We considered vehicle interior and borders to separate instances using a semantic segmentation model. When removing the borders, the vehicle interior becomes isolated, allowing for unique object identification. Our procedure is very efficient and accurate for generating data iteratively, which resulted in 122,567 mapped vehicles. Metrics-wise, our method presented higher IoU when compared to box-based methods (82% against 72%), and per-object metrics surpassed 90% for precision and recall.

show abstract

Vehicle and Vessel Detection on Satellite Imagery: A Comparative Study on Single-Shot Detectors

Cited by 22 publications

References 22 publications

Artificial and beneficial – Exploiting artificial images for aerial vehicle detection

Artificial and beneficial – Exploiting artificial images for aerial vehicle detection

Improving Object Detection in VHR Aerial Orthomosaics

Bounding Box-Free Instance Segmentation Using Semi-Supervised Iterative Learning for Vehicle Detection

Contact Info

Product

Resources

About