Marine animal classification using UMSLI in HBOI optical test facility

Cao, Zheng; Prı́ncipe, José C.; Ouyang, Bing; Dalgleish, Fraser; Vuorenkoski, Anni K.; Ramos, Brian; Alsenas, Gabriel

doi:10.1007/s11042-017-4833-4

Cited by 8 publications

(7 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The experiment for classification is related to DT G policy selecting templates. We select 10 templates by DTG policy and compare its performance with k-means clustering from previous work [20]. Because each class has 256 templates, the previous method calculates 256 × 256 similarity matrix and each row represents one shape template.…”

Section: Classificationmentioning

confidence: 99%

“…Because each class has 256 templates, the previous method calculates 256 × 256 similarity matrix and each row represents one shape template. Then k-means clustering is applied to cluster all templates into 10 categories, the representative of each category is simply selected by the one shape vector that has the least L1-norm [20]. The selection is based on the idea of fully representing the class with fewer instances, which gives reasonably good results for a small data set.…”

Section: Classificationmentioning

confidence: 99%

“…We compare the performance of templates selected by DTG and the original k-means method as well as random policy. The original kmeans method calculates confusion matrix based on previous dataset composed of 8 amberjacks, 6 barracuda and 8 turtles [20]. Because this dataset is too small, we also compare the classification accuracy based on the same dataset for detection task, which only has 2 classes: barracuda and turtle (because templates for other natural fishes we encountered are not available).…”

Section: Classificationmentioning

confidence: 99%

“…10 templates chosen by k-means clustering for each classes. First row: amberjack, second row: barracuda, third row: turtle[20].…”

mentioning

confidence: 99%

See 3 more Smart Citations

Algorithmic Design and Implementation of Unobtrusive Multistatic Serial LiDAR Image

Ding

Cao

Emigh

et al. 2019

Oceans 2019 MTS/Ieee Seattle

Self Cite

View full text Add to dashboard Cite

To fully understand interactions between marine hydrokinetic (MHK) equipment and marine animals, a fast and effective monitoring system is required to capture relevant information whenever underwater animals appear. A new automated underwater imaging system composed of LiDAR (Light Detection and Ranging) imaging hardware and a scene understanding software module named Unobtrusive Multistatic Serial LiDAR Imager (UMSLI) to supervise the presence of animals near turbines. UMSLI integrates the front end LiDAR hardware and a series of software modules to achieve image preprocessing, detection, tracking, segmentation and classification in a hierarchical manner.

show abstract

Section: Classificationmentioning

confidence: 99%

Section: Classificationmentioning

confidence: 99%

Section: Classificationmentioning

confidence: 99%

“…10 templates chosen by k-means clustering for each classes. First row: amberjack, second row: barracuda, third row: turtle[20].…”

mentioning

confidence: 99%

See 2 more Smart Citations

Algorithmic Design and Implementation of Unobtrusive Multistatic Serial LiDAR Image

Ding

Cao

Emigh

et al. 2019

Oceans 2019 MTS/Ieee Seattle

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recently, the Harbor Branch Oceanographic Institute (HBOI) at Florida Atlantic University has developed a novel system called Unobtrusive Multistatic Serial Lidar Imager (UMSLI) to perform marine hydrokinetic site monitoring and marine animal classification [5] [6]. Initial testing of the UMSLI system has been conducted inside a unique test tank facility (Figure 2), which is capable of extensive testing of a variety of different electro-optical system configurations under a range of environmental conditions.…”

mentioning

confidence: 99%

Marine Animal Classification with Correntropy Loss Based Multi-view Learning

Cao¹,

Yu²,

Ouyang³

et al. 2017

Preprint

Self Cite

View full text Add to dashboard Cite

To analyze marine animals behavior, seasonal distribution and abundance, digital imagery can be acquired by visual or Lidar camera. Depending on the quantity and properties of acquired imagery, the animals are characterized as either features (shape, color, texture, etc.), or dissimilarity matrices derived from different shape analysis methods (shape context, internal distance shape context, etc.). For both cases, multi-view learning is critical in integrating more than one set of feature/dissimilarity matrix for higher classification accuracy. This paper adopts correntropy loss as cost function in multi-view learning, which has favorable statistical properties for rejecting noise. For the case of features, the correntropy loss-based multi-view learning and its "entrywise" variation are developed based on the multi-view intact space learning algorithm. For the case of dissimilarity matrices, the robust Euclidean embedding algorithm is extended to its multi-view form with the correntropy loss function. Results from simulated data and real-world marine animal imagery show that the proposed algorithms can effectively enhance classification rate, as well as suppress noise under different noise conditions. I. INTRODUCTIONThe study of marine animals behavior, seasonal distribution and abundance is vital for various environmental agencies, commercial fishermen and marine research institutes. To this end, extensive amount of digital imagery and video are acquired by underwater vehicles (AUVs) and remotely operated vehicles (ROVs). Color images are more intuitive to humans compared sonar Z. Cao, S. Yu and J. C. Principe are with the

show abstract

Underwater De-scattering Imaging by Laser Field Synchronous Scanning

Zhao

2020

Optics and Lasers in Engineering

View full text Add to dashboard Cite

Marine animal classification using UMSLI in HBOI optical test facility

Cited by 8 publications

References 18 publications

Algorithmic Design and Implementation of Unobtrusive Multistatic Serial LiDAR Image

Algorithmic Design and Implementation of Unobtrusive Multistatic Serial LiDAR Image

Marine Animal Classification with Correntropy Loss Based Multi-view Learning

Underwater De-scattering Imaging by Laser Field Synchronous Scanning

Contact Info

Product

Resources

About