| LDR | | 00000nam u2200205 4500 |
| 001 | | 000000434661 |
| 005 | | 20200226165251 |
| 008 | | 200131s2019 ||||||||||||||||| ||eng d |
| 020 | |
▼a 9781687959133 |
| 035 | |
▼a (MiAaPQ)AAI22621829 |
| 040 | |
▼a MiAaPQ
▼c MiAaPQ
▼d 247004 |
| 082 | 0 |
▼a 621.3 |
| 100 | 1 |
▼a Huang, Tsung-Wei. |
| 245 | 10 |
▼a Automatic Video Analysis for Electronic Monitoring of Fishery Activities. |
| 260 | |
▼a [S.l.]:
▼b University of Washington.,
▼c 2019. |
| 260 | 1 |
▼a Ann Arbor:
▼b ProQuest Dissertations & Theses,
▼c 2019. |
| 300 | |
▼a 70 p. |
| 500 | |
▼a Source: Dissertations Abstracts International, Volume: 81-04, Section: B. |
| 500 | |
▼a Advisor: Hwang, Jenq-Neng. |
| 502 | 1 |
▼a Thesis (Ph.D.)--University of Washington, 2019. |
| 506 | |
▼a This item must not be sold to any third party vendors. |
| 506 | |
▼a This item must not be added to any third party search indexes. |
| 520 | |
▼a Recently, automated imagery analysis techniques have drawn increasing attention in fishery science and industry. Compared to traditional human observing and monitoring, automated imagery analysis techniques are more scalable and deployable, and thus have been widely used in recent years for numerous fishery survey tasks, such as abundance estimation, species identification or length measurement. One of the emerging fishery survey tasks which can effectively take advantage of the automated imagery analysis is the electronic monitoring (EM) for fishery activities. The goal of EM is to monitor the fish catching on fishing vessels, either for scientific survey or legal purpose. For example, a fishing vessel may not retain fish catching if the length is below some threshold or exceeding the quota of the vessel for certain species. Therefore, accurate tracking, counting, measurement and species classification is required in the EM systems. There are, however, challenges from the inspected subjects and operation environments. Deformable objects, noise from the wild sea surface, and dynamic background, make conventional tracking, segmentation and classification methods unreliable.To overcome the challenges encountered in the electronic monitoring, this dissertation presents an online 3D tracking and segmentation system for stereo video based monitoring of rail fish catching on wild sea surface. Based on the result of a pre-trained image object (fish) detector, a Kalman filtering-based tracking system overcomes the issues of low detection scores of deformed objects and unreliable bounding boxes by rescoring multiple object proposals using spatial information in 3D. A clustering-and-scoring strategy is then applied on the depth map, so that a plane classification method can effectively segment the objects from the dynamic background without any prior modeling. The object segmentation is further refined using fully connected conditional random fields based on color and geometric features. With the segmentation results, we can measure the 3D lengths of objects and update the positions of bounding boxes to help tracking. Experimental results show that a reliable tracking and measurement performance under noisy and dynamic sea surface environment is achieved.Once fish are tracked and measured, one of the primary tasks in the electronic monitoring is to identify the species of fish. In the work of object classification, one challenge is that the feature generation needs to be robust with fish in any orientations and poses, which yield diverse visual features and large within-class variation. Other challenges include the high visual similarity among fish species. Therefore, in this dissertation, we utilize the deep metric learning to learn a feature representation which can separate visually similar species in the feature space. By adding more constraints based on the temporal order of image sequences, we can make the model to learn a more structured and compact feature space. Besides, by exploiting the clustering properties and temporal relationship in the feature space, the learning of the model can be further improved.Combining all the stages above, the proposed electronic monitoring system can automatically process input video data, and analyze the fishery activities. The monitoring results can be further used to perform fish abundance estimation for either regulatory or scientific research purposes. |
| 590 | |
▼a School code: 0250. |
| 650 | 4 |
▼a Electrical engineering. |
| 690 | |
▼a 0544 |
| 710 | 20 |
▼a University of Washington.
▼b Electrical Engineering. |
| 773 | 0 |
▼t Dissertations Abstracts International
▼g 81-04B. |
| 773 | |
▼t Dissertation Abstract International |
| 790 | |
▼a 0250 |
| 791 | |
▼a Ph.D. |
| 792 | |
▼a 2019 |
| 793 | |
▼a English |
| 856 | 40 |
▼u http://www.riss.kr/pdu/ddodLink.do?id=T15493847
▼n KERIS
▼z 이 자료의 원문은 한국교육학술정보원에서 제공합니다. |
| 980 | |
▼a 202002
▼f 2020 |
| 990 | |
▼a ***1008102 |
| 991 | |
▼a E-BOOK |