Wyniki wyszukiwania - BazTech

1

Selection of an artificial pre-training neural network for the classification of inland vessels based on their images

Bobkowska Katarzyna, Bodus-Olkowska Izabela

Zeszyty Naukowe Akademii Morskiej w Szczecinie

|

2021

|

nr 67 (139)

91--97

EN

Artificial neural networks (ANN) are the most commonly used algorithms for image classification problems. An image classifier takes an image or video as input and classifies it into one of the possible categories that it was trained to identify. They are applied in various areas such as security, defense, healthcare, biology, forensics, communication, etc. There is no need to create one’s own ANN because there are several pre-trained networks already available. The aim of the SHREC projects (automatic ship recognition and identification) is to classify and identify the vessels based on images obtained from closed-circuit television (CCTV) cameras. For this purpose, a dataset of vessel images was collected during 2018, 2019, and 2020 video measurement campaigns. The authors of this article used three pre-trained neural networks, GoogLeNet, AlexNet, and SqeezeNet, to examine the classification possibility and assess its quality. About 8000 vessel images were used, which were categorized into seven categories: barge, special-purpose service ships, motor yachts with a motorboat, passenger ships, sailing yachts, kayaks, and others. A comparison of the results using neural networks to classify floating inland units is presented.

2

Potential and use of the googlenet ann for the purposes of inland water ships classification

Bobkowska Katarzyna, Bodus-Olkowska Izabela

Polish Maritime Research

|

2020

|

nr 4

170--178

EN

This article presents an analysis of the possibilities of using the pre-degraded GoogLeNet artificial neural network to classify inland vessels. Inland water authorities monitor the intensity of the vessels via CCTV. Such classification seems to be an improvement in their statutory tasks. The automatic classification of the inland vessels from video recording is a one of the main objectives of the Automatic Ship Recognition and Identification (SHREC) project. The image repository for the training purposes consists about 6,000 images of different categories of the vessels. Some images were gathered from internet websites, and some were collected by the project’s video cameras. The GoogLeNet network was trained and tested using 11 variants. These variants assumed modifications of image sets representing (e.g., change in the number of classes, change of class types, initial reconstruction of images, removal of images of insufficient quality). The final result of the classification quality was 83.6%. The newly obtained neural network can be an extension and a component of a comprehensive geoinformatics system for vessel recognition.

3

The Hough transform in the classification process of inland ships

Bobkowska Katarzyna, Wawrzyniak Natalia

Zeszyty Naukowe Akademii Morskiej w Szczecinie

|

2019

|

nr 58 (130)

9--15

EN

This article presents an analysis of the possibilities of using image processing methods for feature extraction that allows kNN classification based on a ship’s image delivered from an on-water video surveillance system. The subject of the analysis is the Hough transform which enables the detection of straight lines in an image. The recognized straight lines and the information about them serve as features in the classification process. Above all, this approach allows ships to be recognized, which can then be characterized by a specific representation and shape. Recreational units that are often seen on inland waters were classified correctly using this method. Each analyzed camera image was previously prepared – brought to the form where the ship was visible from the side and the background removed (they were monochromatic – white). The results obtained in this work will allow for the development of the final ship classification method based on camera images. This method is a significant part of the emerging system prototype, which is implemented as part of the Automatic Ship Recognition and Identification (SHREC) project.