Unlock instant, AI-driven research and patent intelligence for your innovation.

Classification method for training sample expansion according to recognition results of multiple classifiers

A classification method and training sample technology, applied in the classification field of training sample expansion, can solve problems such as time-consuming, and achieve the effects of reducing interference, improving classification accuracy, and improving classification speed.

Pending Publication Date: 2022-07-05
GUANGXI ZHUANG AUTONOMOUS REGION ACAD OF AGRI SCI
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

At present, the multi-classifier fusion algorithm mainly uses the results of multiple classifiers for weighted or non-weighted voting, and the final classification result is formed by the voting results. Compared with the single classifier method, the classification accuracy is usually higher, but it takes longer

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
  • Classification method for training sample expansion according to recognition results of multiple classifiers
  • Classification method for training sample expansion according to recognition results of multiple classifiers
  • Classification method for training sample expansion according to recognition results of multiple classifiers

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0027] see figure 1 , the classification method for expanding training samples according to the multi-classifier identification results in this embodiment includes the following contents:

[0028] Step S10. Preparation procedure: select the classification methods used for various classifications; select the initial training samples including each category in the data set to be classified, and form the initial training sample set; wherein, the classification method and the training sample set before and after the preparation sequence Can be exchanged with each other as needed, or prepared at the same time;

[0029] Step S20. Classification and identification program: use each classification method and training sample set to classify and identify the data of the data set to be classified, and obtain the classification results of the data set to be classified by each classification method;

[0030] Step S30. Approval rate calculation program: According to the classification resu...

Embodiment 2

[0057] Based on the foregoing embodiment 1, the present embodiment 2 provides an example of rice canopy coverage classification, which is as follows:

[0058] Canopy cover is a useful indicator for assessing crop growth and predicting crop yield. Usually, a digital camera can take pictures of crop canopy, use classification algorithm to classify and identify the pictures, divide different types of areas in the image, obtain the range of leaf canopy in the picture and calculate its proportion, then the canopy coverage of crops can be obtained. .

[0059] Image of the rice canopy taken with a digital camera in this case (see Figure 4 ) as the data to be classified, the size of the image is 850*650, a total of 552,500 pixels, and there are 3 layers of R, G, and B. Use the aforementioned classification method to classify the image, and identify four categories of rice canopy leaves, water body, soil, and shadow in the image. See the identification of each category. Figure 5-F...

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 discloses a classification method for training sample expansion according to recognition results of multiple classifiers. The classification method comprises the following steps: selecting a plurality of classification methods for classification; selecting initial training samples including various categories from the to-be-classified data set to form an initial training sample set; classifying and identifying the data of the to-be-classified data set by adopting each classification method and the training sample set to obtain a to-be-classified data set classification result of each classification method; according to the classification result of the to-be-classified data set, calculating the classification result recognition rate of each piece of data; comparing the classification result recognition rate with a preset threshold value to obtain a new training sample, and expanding a training sample set; according to the expanded training sample set, judging and executing next iteration classification; and taking the maximum approved classification result of the last iterative operation as the final data classification result of the to-be-classified data set. According to the invention, through mutual iteration verification of multi-classifier results, training samples are gradually amplified, and the classification precision is improved.

Description

technical field [0001] The invention relates to the technical field of multi-classifier fusion algorithms, in particular to a classification method for expanding training samples according to multi-classifier identification results. Background technique [0002] The function of the classifier is to judge the category to which a new observation sample belongs based on the training data of the labeled category. The conventional task of the classifier is to use a given category and known training data to learn classification rules, and then classify (or predict) the unknown data. It is widely used in many fields such as target recognition, weather forecast, signal processing, and image classification. , scholars have proposed a variety of mature classifiers. The differences in the classification performance of different classifiers are related to the statistical distribution characteristics of the classified data, prior knowledge, the size of the training data samples, and the...

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): G06K9/62
CPCG06F18/241G06F18/214
Inventor 杨绍锷黄启厅谭黎光谢国雪谭序光
Owner GUANGXI ZHUANG AUTONOMOUS REGION ACAD OF AGRI SCI