A method for extracting painted pottery patterns based on deep learning
A deep learning, deep learning network technology, applied in hyperspectral image processing, deep learning-based extraction of painted pottery patterns, can solve problems such as poor fusion effect
Active Publication Date: 2021-09-17
NORTHWEST UNIV
View PDF6 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] Aiming at the problem in the prior art that the acquired principal component images of painted pottery contain a lot of noise, resulting in poor subsequent fusion effects, the present invention provides a method for extracting painted pottery patterns based on deep learning, including the following steps:
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0066] Step 1: Obtain the painted pottery data collected in November 2016 at the College of Arts and Sciences, Northwest University as experimental data;
[0067] Step 2: Perform image cropping, band deletion, and radiation correction preprocessing on the collected data. The preprocessed data artificially selects the representative bands of visible light (red: 645nm, green: 550nm, blue: 479nm) respectively corresponding to three RGB channel, to synthesize a true-color image, such as image 3 (b).
[0068] Step 3: Do minimum noise separation on the preprocessed data to obtain a principal component image containing pattern information, such as image 3 (c).
[0069] Step 4: Right image 3 (c) Do binarization processing to obtain binarized image 3(d), use the above method according to image 3 (c) and image 3 (d) Obtain training samples and labels, and train the deep learning network; image 3 (c) As the input of the trained network, obtain the pattern information image, s...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
A method for extracting painted pottery patterns based on deep learning, comprising: step 1: obtaining the original hyperspectral data of painted pottery, and preprocessing it; step 2: synthesizing a true-color image according to the image obtained in step 1; step 3: Perform minimum noise separation processing on the image obtained in step 1 to obtain a principal component image containing pattern information; step 4: perform binarization processing on the image obtained in step 3, obtain training samples and training sample labels, and input them into the deep learning network to obtain Output pattern information image; step 5: extract detail information through image decomposition based on sparse representation, and inject detail information into the true-color image in step 2 through the detail injection model to recover the pattern information of painted pottery. The invention extracts patterns from principal component images, recovers a small amount of lost information while extracting clear patterns, and lays a solid foundation for the research and protection of painted cultural relics.
Description
technical field [0001] The invention belongs to the field of information technology and relates to hyperspectral image processing technology, in particular to a method for extracting painted pottery patterns based on deep learning. Background technique [0002] Painted pottery often has exquisite patterns and rich pigments on the surface, which is a precious cultural relic with rich historical and cultural value. However, painted pottery was buried in the ground for a long time, and most of the patterns and pigments were lost or covered by earth. Under visible light, it is difficult for people to observe pattern information with the naked eye. [0003] In order to obtain clearer and more accurate pattern information, people often use hyperspectral imaging technology to protect cultural relics. As a comprehensive technology, hyperspectral imaging technology has the advantage of not causing damage to the data itself in the process of collecting data, and hyperspectral imagin...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/12G06T5/00G06N3/08
CPCG06N3/08G06T7/12G06T2207/20081G06T5/70
Inventor 王珺俞凯彭进业祝轩刘成李展章勇勤罗迒哉王琳樊萍
Owner NORTHWEST UNIV