Unstructured feature differentiation method based on multi-intelligence fusion

An unstructured and intelligent technology, applied in image analysis, color/spectral characteristic measurement, image data processing, etc., can solve the problem of low accuracy of identification results of traditional Chinese medicine, single characteristics of medicinal materials, and poor extraction of shape and texture features. problems, to broaden the dimension and scale of identification, improve accuracy, and solve the effects of difficult identification

Inactive Publication Date: 2018-06-29
NANJING UNIVERSITY OF TRADITIONAL CHINESE MEDICINE
View PDF5 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] 1. The commercially available visual bionic system only analyzes the color features of Chinese herbal medicines, and does not perform well in extracting shape and texture features
And its principle is mainly based on the analysis of the chromaticity difference of the color, so the "dyeing and counterfeiting" of Chinese herbal medicines cannot be effectively identified;
[0005] 2. The olfactory bionic system is only sensitive to the volatile components in the smell of Chinese medicinal materials, and the characteristics of the extracted medicinal materials are too single;
[0006] 3. The taste bionic system is only suitable for the determination of dissolved organic and inorganic components in liquid samples
[0007] The trait information of Chinese medicinal materials obtained by the three methods is somewhat one-sided, which reduces the accuracy of the identification results of Chinese medicinal materials and is not suitable for large-scale promotion.

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Unstructured feature differentiation method based on multi-intelligence fusion
  • Unstructured feature differentiation method based on multi-intelligence fusion
  • Unstructured feature differentiation method based on multi-intelligence fusion

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0028] Using ultra-large-scale feedback neural network pattern intelligent recognition technology to extract and analyze the shape, texture, and surface vector features of medicinal materials.

[0029] (1) Take 20 pieces of dried licorice decoction pieces, and use a camera to obtain images. The image is binarized to clearly reflect the target from the background. Denoise the binarized image, extract the feature vector and then classify it.

[0030] (2) The binarization processing operation process in step (1) is: set the threshold T, the pixel group greater than T takes the value 1, and the pixel group smaller than T takes the value 0. The binarization function is as follows:

[0031]

[0032] (3) In step (1), use the following statement to remove noise: x4=medfilt2(x3,[10,10])

[0033] (4) Step (1) is graded according to the following criteria:

[0034] Table 1 Sample color characteristic value standard

[0035] Herb Grade

Hue (hue, H)

Saturation (sat...

Embodiment 2

[0039] An array gas signal sensor is used to extract and analyze the vector characteristics of the odor components of medicinal materials.

[0040] (1) Take 20 dried licorice pieces, grind them into powder and pass through a No. 2 sieve, take 1.0 g of sample powder, put it in a 20 mL headspace sampler, and place it in an array-type intelligent signal extraction system. Set the shaking time to 300s, and the lowest shaking temperature to 35°C. The odor recognition model of licorice decoction pieces was established by principal component analysis (PCA).

[0041] (2) In step (1), the PCA analysis method can be obtained as figure 1 analysis chart. Area 1 in the figure is the odor range of first-class licorice decoction pieces, area 2 is the odor range of second-class licorice decoction pieces, and area 3 is the odor range of a sample. The size of the distance of each area in the figure reflects the closeness of each smell, so it can be judged that the sample is a second-class pr...

Embodiment 3

[0043] The complex component multi-channel discrimination intelligent system is used to extract and analyze the vector characteristics of the taste components of medicinal materials.

[0044] (1) Take 20 pieces of dried licorice decoction pieces, powder them through a No. 3 sieve, take 1.0g of sample powder in a 250mL conical flask, add 80mL of water, soak for 30min, heat and reflux for 1h, let cool, filter, take 20mL of the filtrate and put it in a complex Component multi-channel discrimination and analysis intelligent system. The taste recognition model of licorice decoction pieces was established by principal component analysis (PCA).

[0045] (2) In step (1), the PCA analysis method can be obtained as figure 2 analysis chart. Area 1 in the figure is the taste range of first-class licorice decoction pieces, area 2 is the taste range of second-class licorice decoction pieces, and area 3 is the taste range of a sample. The size of the distance between each area in the fig...

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

No PUM Login to view more

Abstract

The invention relates to the technical field of automation, and discloses an unstructured feature differentiation method based on multi-intelligence fusion which is used for identifying properties oftraditional Chinese medicines. By combining an super-large-scale feedback type neural network mode intelligent identification system, an array type intelligent signal extraction system and a multi-channel intelligent complex ingredient identification system, the method can simultaneously effectively extract color features, size features, texture features, odor features and taste features of traditional Chinese medicines and comprehensively analyze the features, so that the specifications and grades of the traditional Chinese medicines can be accurately differentiated.

Description

technical field [0001] The invention relates to a method for distinguishing unstructured features, in particular to a method for distinguishing unstructured features based on the fusion of multiple intelligences and its application in character identification of traditional Chinese medicines. Background technique [0002] In recent years, with the development of micro-manufacturing technology and micro-electronic machining technology, bionic systems for vision, smell, and taste have come out one after another, and have been tried in the fields of food, environment, and medicine, and can objectively evaluate certain traits. [0003] But there are following problems: [0004] 1. The commercially available visual bionic system only analyzes the color features of Chinese herbal medicines, and does not perform well in extracting shape and texture features. And its principle is mainly based on the analysis of the chromaticity difference of the color, so the "dyeing and counterfei...

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
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G01N21/25
CPCG06T7/90G01N21/25G06T2207/20084G06T2207/10024G06F18/2411
Inventor 徐飞杜文嘉陆彩尤敏
Owner NANJING UNIVERSITY OF TRADITIONAL CHINESE MEDICINE
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap