Abstract
Recent work has shown that recognizing point symbols is an essential task in the field of map digitization. For the identification of symbols, it is generally necessary to compare the symbols with a specific criterion and find the most similar one with each known symbol one by one. Most of the works can only identify a single symbol, a small number of works are to deal with multiple symbols simultaneously with a low recognition accuracy. Given the two deficiencies, this paper proposes a deep transfer learning architecture, where the task is to learn a symbol classifier with AlexNet. For the insufficient dataset, we develop a method for transfer learning that uses a MNIST dataset to pretrain the model, which makes up for the problem of small training dataset and enhances the generalization of the model. Before the recognition process, preprocessing the point symbols in the map to coarse screening out the areas suspected of point symbols. We show a significant improvement over using point symbol images to keep a high performance in being able to deal with many more categories of symbols simultaneously.
Highlights
In resource development, engineering construction, and military defense [1], nowadays, many scientific research institutions and researchers have proposed some algorithms for the recognition of point symbols in color topographic maps.Recognizing a map consisting of points, lines, and surface symbols is equivalent to identifying various points, lines, and surface symbols
Since the line symbols are composed of basic point symbols through multiple combinations, automatic recognition of the point symbols is the core for the entire map symbol recognition
The recognition method with the deep transfer learning improves the accuracy of recognition and handles multiple point symbols simultaneously
Summary
In resource development, engineering construction, and military defense [1], nowadays, many scientific research institutions and researchers have proposed some algorithms for the recognition of point symbols in color topographic maps. An interesting method [10] has been proposed by decomposing each pixel and the spatial neighborhood into a low-rank form, and the spatial information can be efficiently integrated into the spectral signatures This way relies relatively on symbol extraction, for the complex topographic features, the extraction results are inevitably less than satisfactory. The improved back-propagation(BP) neural network is introduced to identify point symbols by some researchers at present [15]. The recognition method with the deep transfer learning improves the accuracy of recognition and handles multiple point symbols simultaneously. (1) We experimentally show that the method of deep neural networks is good feature learning and classifying machines that model recognizes multi-symbols simultaneously.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
