Transgenic crop detection method combining terahertz spectrum and mean shift algorithm

Abstract

The invention discloses a transgenic crop detection method combining a terahertz spectrum and a mean shift algorithm, and belongs to the field of molecular biology. The method comprises the followingsteps: preparing a to-be-detected sample into a solid tabletting sample; performing terahertz time-domain spectrum scanning on the sample to obtain a sample terahertz absorption spectrum; and analyzing the terahertz absorption spectrum of the sample by using a PCA algorithm and a Mean Shift algorithm. According to the method, principal component analysis and a mean shift algorithm are combined; when different varieties of crops exist in the sample, the crops can be automatically distinguished into different categories on the premise that a training set does not need to be constructed; when whether a certain sample is a transgenic crop or not needs to be specifically identified in one step, only the standard database is needed for comparison. According to the method, the problems that a training set and a test set need to be constructed and samples of the training set need to be manually labeled are avoided in the whole process; the method is suitable for detection of a small number ofsamples, and reduces the manual workload and intervention errors.

Description

technical field [0001] The invention relates to the technical field of molecular biology, in particular to a transgenic crop detection method combining terahertz spectrum and mean shift algorithm. Background technique [0002] Genetically modified crops refer to the use of genetic engineering to add the genes of the original crops to the genetic material of other organisms and remove the bad genes, resulting in better quality crops. Since the commercialization of genetically modified crops, the planting area of ​​genetically modified crops in the world has shown a rapid growth trend. At present, the planting varieties of genetically modified crops are mainly four types of crops: soybean, corn, cotton and rapeseed. According to relevant statistics, as of 2017, the global planting area of ​​genetically modified crops reached 189.8 million hectares. Although compared with traditional crops, genetically modified crops have the advantages of higher yield, drought resistance and...

AI Technical Summary

Problems solved by technology

When the number of samples collected is small, the number of samples divided into the training set will be reduced accordingly, which will lead to overfitting problems in the model during training, affect the prediction performance of the model, and even make it impossible to distinguish whether it is a genetically modified crop

[0011] Large manual workload: due to the need to manually label the samples in the training set when constructing the training set

When the sample size in the training set is large, more workload is required, and at the same time, it is easy to cause errors in training set construction due to manual errors.

Images