Saliency-based classification of objects in unconstrained underwater environments
- PDF / 973,165 Bytes
- 17 Pages / 439.37 x 666.142 pts Page_size
- 94 Downloads / 194 Views
Saliency-based classification of objects in unconstrained underwater environments Nitin Kumar 1 & H. K. Sardana 1,2
& S. N. Shome
1,3
& Vishavpreet Singh
2
Received: 9 August 2019 / Revised: 3 June 2020 / Accepted: 15 June 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
Exploration of the deep-sea underwater environment is a challenging and non-trivial task. Underwater vehicles used for the exploration of such environments capture videos continuously. The processing of these videos is a major bottleneck for scientific research in this area. This paper presents a methodology for the classification of the objects in the unconstrained underwater environments into two broad classes namely - man-made and natural. The classification of the objects is achieved using the saliency gradient based morphological active contour models. A bag of features acquired from the contours of the objects is used for the classification using various classifiers. Principal Component Analysis is used for the removal of redundancy in the feature set. The proposed features classify the objects present in the unconstrained underwater environment into a manmade/natural class using the proposed features. The results show that all the classifiers performed well; though KNN and ensemble subspace KNN, performed marginally better. Keywords Classification . Bag of features . SVM . KNN . Ensemble subspace KNN . Saliency gradient based morphological active contour models
* H. K. Sardana [email protected] Nitin Kumar [email protected] S. N. Shome [email protected] Vishavpreet Singh [email protected]
1
Academy of Scientific and Innovative Research (AcSIR), Ghaziabad, India
2
CSIR-Central Scientific Instruments Organisation (CSIR-CSIO), Chandigarh, India
3
CSIR-Central Mechanical Engineering Research Institute (CSIR-CMERI), Durgapur, West Bengal, India
Multimedia Tools and Applications
1 Introduction Exploration of the underwater space is of great importance for meeting the ever-increasing demand for resources on earth. Sonar based methods [1–3], address the problem of underwater environment exploration, however, for limited range detection, optical imaging plays a vital role in underwater environments. During the last two decades, extensive research is being carried out in the field of underwater vehicles for exploring the deep-sea environments. These vehicles act as platforms for attaching various sensors. A large number of videos are being captured continuously by the cameras mounted on these vehicles [4]. The processing of such a large amount of videos manually is a major bottleneck for the underwater research community. Automated processing of these videos is the need of the hour for the exploration of the underwater space. Underwater imaging is a challenging field for observing the things happening therein. Imaging in underwater environment is challenging as properties of light changes on entering the water medium and varies with depth. Moreover, light scattering and absor
Data Loading...
