Computer Vision and Applications: A Guide for Students and Practitioners

Abstract. We present a monitoring technique tailored to analysing change from near-continuously collected, high-resolution 3-D data. Our aim is to fully characterise geomorphological change typified by an event magnitude-frequency relationship that adheres to an inverse power law or similar. While recent advances in monitoring have enabled changes in volume across more than 7 orders of magnitude to be captured, event frequency is commonly assumed to be interchangeable with the time-averaged event numbers between successive surveys. Where events coincide, or coalesce, or where the mechanisms driving change are not spatially independent, apparent event frequency must be partially determined by survey interval.The data reported have been obtained from a permanently installed terrestrial laser scanner, which permits an increased frequency of surveys. Surveying from a single position raises challenges, given the single viewpoint onto a complex surface and the need for computational efficiency associated with handling a large time series of 3-D data. A workflow is presented that optimises the detection of change by filtering and aligning scans to improve repeatability. An adaptation of the M3C2 algorithm is used to detect 3-D change to overcome data inconsistencies between scans. Individual rockfall geometries are then extracted and the associated volumetric errors modelled. The utility of this approach is demonstrated using a dataset of ∼ 9 × 10 3 surveys acquired at ∼ 1 h intervals over 10 months. The magnitude-frequency distribution of rockfall volumes generated is shown to be sensitive to monitoring frequency. Using a 1 h interval between surveys, rather than 30 days, the volume contribution from small (< 0.1 m 3 ) rockfalls increases from 67 to 98 % of the total, and the number of individual rockfalls observed increases by over 3 orders of magnitude. High-frequency monitoring therefore holds considerable implications for magnitude-frequency derivatives, such as hazard return intervals and erosion rates. As such, while high-frequency monitoring has potential to describe short-term controls on geomorphological change and more realistic magnitude-frequency relationships, the assessment of longer-term erosion rates may be more suited to less-frequent data collection with lower accumulative errors.

show abstract

“…This threshold, however, represents the maximum possible error associated with the rockfall area. Jahne (2000) defined the variance σ 2…”

Section: Extracting Discrete Changes and Quantifying Volumetric Errormentioning

confidence: 99%

Optimising 4-D surface change detection: an approach for capturing rockfall magnitude–frequency

et al. 2018

View full text Add to dashboard Cite

show abstract

“…Blobs obtained from the detection section are tracked using region correspondence [10]. Parameters such as centroid, bounding box, size, velocity and change in size of each blob are extracted.…”

Section: Object Trackingmentioning

confidence: 99%

A New RMI Framework for Outdoor Objects Recognition

Ern

Joo

2009

2009 International Conference on Advanced Computer Control

View full text Add to dashboard Cite

show abstract

“…Jahne and Haußecker defined computer vision as "a set of techniques for the collection, processing, analysis and understanding of complex multidimensional data from our environment for technical and scientific applications" (Jahne et al, 2000). A lot of algorithms have been developed in the field of computer vision and one subgroup is used to remove background from video streams.…”

Section: Computer Visionmentioning

confidence: 99%

Model of information density measuring in e-learning videos

Đokić¹,

Idlbek²,

Mandušić³

2017

PROSOC

View full text Add to dashboard Cite

Educational video content authors must be careful that the content is adequate for the audience and the purpose especially on the term level. On the other hand authors must be careful that the content does not cause information overloading. Viewers can experience it if the videos have too much motions and movements. Information overload has become an increasingly popular area of study and a lot of authors indicated that information density of movies and videos have risen in the last few decades. High information density of videos and movies may cause information overloading and it has been proved by Lang et al in their theoretic model named Limited Capacity Model of Motivated Mediated Message Processing. This theoretic model suggests how to measure information density in videos and movies but this method is slow and just trained observers can do this job well. In this paper the new model for information density measuring in videos and movies is described and tested. This new model is based on special usage of background subtraction algorithms and in this paper multi-layer background subtraction algorithm based on color and texture is used. This algorithm is presented by authors Yao and Odobez on CVPR Visual Surveillance Workshop in 2007. Results of our model are compared with the trained observer's results and the same video clips that have been used by Lang et al are used for testing. It can be seen that there is a significant correlation between information densities measured by trained observers and our model. Our model can be used to determine some forms of information density in videos and movies and it can be used to indicate higher possibility of information overloading.

show abstract

Computer Vision and Applications: A Guide for Students and Practitioners

Cited by 52 publications

References 0 publications

Optimising 4-D surface change detection: an approach for capturing rockfall magnitude–frequency

Optimising 4-D surface change detection: an approach for capturing rockfall magnitude–frequency

A New RMI Framework for Outdoor Objects Recognition

Model of information density measuring in e-learning videos

Contact Info

Product

Resources

About