Manhole Cover Detection on Rasterized Mobile Mapping Point Cloud Data Using Transfer Learned Fully Convolutional Neural Networks

Mattheuwsen, L.; Vergauwen, Maarten

doi:10.3390/rs12223820

Cited by 18 publications

(7 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Additionally, the marked point approach from (Yu et al, 2014) is used to accurately determine the final location and dimensions of the manhole covers. A similar approach on intensity ground images was proposed in our previous work (Mattheuwsen and Vergauwen, 2020), where we transfer learned several different backbone architectures on a relatively small training dataset to detect manhole covers. Our approach achieves 97.3% recall and 97.3% precision on the road surface and is able to locate the storm drains with a 2D confidence interval of 16.5 cm.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Storm Drain Detection and Localisation on Mobile Lidar Data Using a Pre-Trained Randla-Net Semantic Segmentation Network

Mattheuwsen

Bassier

Vergauwen

2022

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. As the expansion of cities and urban areas results in the construction of more impermeable road surfaces, a well designed urban drainage system becomes of greater importance. However, the accurate and up-to-date mapping of storm drains necessary to create accurate drainage models is often lacking. In recent years, mapping of the road infrastructure is increasingly carried out by highly efficient mobile mapping systems but which lack automatic interpretation of the massive amount data. In this paper we present a fully automatic storm drain detection method to extract and locate storm drain inlets in mobile mapping lidar data. The point cloud is first segmented by a pre-trained RandLa-Net model, which although untrained to segment storm drains, is able the segment storm drain clusters in the hardscape class. The results from this class are further processed by enforcing different requirements to only extract and locate storm drain clusters. Our approach is evaluated on a large testing dataset with 171 storm drains and achieves 81.9%, 95.2% and 88.1% for recall, precision and F1-score respectively. The majority of the false positive and false negative detections are due to incorrect point cloud segmentation of the RandLa-Net. In terms of localisation, our approach achieves an RMSE of 5.5 cm on the centre location while the dimensions of the bounding box are on average 23% off compared to the ground truth.

show abstract

Section: Related Workmentioning

confidence: 99%

“…However, with the 2D projection of the point cloud, the vertical information of the data is lost. In our previous work (Mattheuwsen and Vergauwen, 2020), we investigated an approach which captured the vertical information within a channel of the intensity ground image, but were unsuccessful in improving the results.…”

Section: Related Workmentioning

confidence: 99%

Storm Drain Detection and Localisation on Mobile Lidar Data Using a Pre-Trained Randla-Net Semantic Segmentation Network

Mattheuwsen

Bassier

Vergauwen

2022

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…Thus, to efficiently map shortcuts on larger scales, automated algorithms for shaft localization, using remote sensing data, could be used (e.g. Mattheuwsen and Vergauwen, 2020;Moy de Vitry et al, 2018). An application of the local connectivity model to larger scales could then replace the extrapolation approach used in this study, eliminating the associated uncertainty.…”

Section: Further Researchmentioning

confidence: 99%

Hydraulic shortcuts increase the connectivity of arable land areas to surface waters

Schönenberger

Stamm

2021

Hydrol. Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract. Surface runoff represents a major pathway for pesticide transport from agricultural areas to surface waters. The influence of artificial structures (e.g. roads, hedges, and ditches) on surface runoff connectivity has been shown in various studies. In Switzerland, so-called hydraulic shortcuts (e.g. inlet and maintenance shafts of road or field storm drainage systems) have been shown to influence surface runoff connectivity and related pesticide transport. Their occurrence and their influence on surface runoff and pesticide connectivity have, however, not been studied systematically. To address that deficit, we randomly selected 20 study areas (average size of 3.5 km2) throughout the Swiss plateau, representing arable cropping systems. We assessed shortcut occurrence in these study areas using three mapping methods, namely field mapping, drainage plans, and high-resolution aerial images. Surface runoff connectivity in the study areas was analysed using a 2×2 m digital elevation model and a multiple-flow algorithm. Parameter uncertainty affecting this analysis was addressed by a Monte Carlo simulation. With our approach, agricultural areas were divided into areas that are either directly, indirectly (i.e. via hydraulic shortcuts), or not at all connected to surface waters. Finally, the results of this connectivity analysis were scaled up to the national level, using a regression model based on topographic descriptors, and were then compared to an existing national connectivity model. Inlet shafts of the road storm drainage system were identified as the main shortcuts. On average, we found 0.84 inlet shafts and a total of 2.0 shafts per hectare of agricultural land. In the study catchments, between 43 % and 74 % of the agricultural area is connected to surface waters via hydraulic shortcuts. On the national level, this fraction is similar and lies between 47 % and 60 %. Considering our empirical observations led to shifts in estimated fractions of connected areas compared to the previous connectivity model. The differences were most pronounced in flat areas of river valleys. These numbers suggest that transport through hydraulic shortcuts is an important pesticide flow path in a landscape where many engineered structures exist to drain excess water from fields and roads. However, this transport process is currently not considered in Swiss pesticide legislation and authorization. Therefore, current regulations may fall short in addressing the full extent of the pesticide problem. However, independent measurements of water flow and pesticide transport to quantify the contribution of shortcuts and validating the model results are lacking. Overall, the findings highlight the relevance of better understanding the connectivity between fields and receiving waters and the underlying factors and physical structures in the landscape.

show abstract

“…The average efficiency of highlighting columnar objects is 87%. Automatic method for detection of hatch covers is proposed in article [4]. The cloud of mobile mapping points is taken as the basis.…”

Section: Introductionmentioning

confidence: 99%

Image processing for automatic road inventory

Белим

Khiryanov

Kvashnina

et al. 2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

Road image analysis is an important task for automatic road inventory. The determination geometric dimensions for the road and the identification road objects are subprocess of constructing a road digital image. In this article, two algorithms for solving different subtasks of automatic road image inventory are proposed. The first algorithm identifies road signs. A convolutional artificial neural network is used in this algorithm. The training set for the neural network is prepared. A computer experiment to determine the recognition effectiveness of road signs has been conducted. The second algorithm defines the edges of the pavement. The algorithm consists five stages. The edges of the road are modeled as straight lines. The result allows you to automatically determine the width of the road.

show abstract

Manhole Cover Detection on Rasterized Mobile Mapping Point Cloud Data Using Transfer Learned Fully Convolutional Neural Networks

Cited by 18 publications

References 28 publications

Storm Drain Detection and Localisation on Mobile Lidar Data Using a Pre-Trained Randla-Net Semantic Segmentation Network

Storm Drain Detection and Localisation on Mobile Lidar Data Using a Pre-Trained Randla-Net Semantic Segmentation Network

Hydraulic shortcuts increase the connectivity of arable land areas to surface waters

Image processing for automatic road inventory

Contact Info

Product

Resources

About