A method for image segmentation of ships in sea and sky background based on joint image information
A ship image and image information technology, applied in ship image segmentation, ship image segmentation in the sea-sky background based on joint image information, can solve the problem of reduced segmentation accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0071] This embodiment is a method for segmenting ships in the sea-sky background based on joint image information; in this embodiment and examples, part of the data set used is from the Singapore Maritime Dataset (SMD), and the other part comes from the RoCoCo experiment of the University of Rome The Maritime Detection, Classification, and Tracking Dataset (MarDCT) created by the laboratory, the remaining small part comes from the network.
[0072] The ship segmentation method in the sea-sky background based on joint image information includes the following steps:
[0073] Step a: Preprocess the image:
[0074] Perform normalization and image scaling operations on all images;
[0075] The purpose of the normalization operation is to prevent the saturation of the neuron output due to the excessively large absolute value of the net input;
[0076] The image scaling operation is used because the inconsistent image size is not friendly to training neural networks. In this embo...
Embodiment 1
[0106] In this embodiment, the present invention is described by taking a ship image under a normal sea and sky background and a ship image under sea fog as examples.
[0107] Specifically, follow these steps:
[0108] 1. Data preparation stage
[0109] All the data used for training and testing used in this embodiment 1 come from SMD, MarDCT (this data set is well known to those skilled in the art, and is not repeated in this embodiment of the present invention) and the network. All data were manually labeled using LabelMe. The reorganized ship database was later called Mari-Data. Since the present invention needs to train the interference factor discriminator and the ship extractor separately, the training data set is also divided into two parts.
[0110] For the data of the interference factor discriminator, this embodiment only considers the sea fog scene that has the most serious impact on the accuracy, thereby improving the overall segmentation accuracy. Therefore, th...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com