Biochemical substance terahertz spectrum identification method based on YOLO network

CN119131411APending Publication Date: 2024-12-13PEKING UNIV
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411219496.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-02
Publication Date
2024-12-13

AI Technical Summary

Technical Problem

The existing spectral recognition methods work hard, time-consuming and low recognition accuracy when identifying the terahertz spectrum of biochemical substances, and it is particularly difficult to identify the terahertz spectrum data of fragmented biochemical substances.

Method used

Using the YOLO network-based method, the terahertz spectrum of biochemical substances is obtained through the terahertz time domain spectroscopy system, absorption peak annotation and data set segmentation, and the YOLO v5 model is constructed and trained to identify the terahertz spectrum of biochemical substances.

Benefits of technology

It improves the recognition accuracy of the terahertz spectrum of biochemical substances, can effectively identify the entire or fragmented terahertz spectrum images, and provides an efficient identification method.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119131411A_ABST
    Figure CN119131411A_ABST
Patent Text Reader

Abstract

The invention discloses a biochemical substance terahertz spectrum recognition method based on a YOLO network, and belongs to the field of terahertz spectrum recognition. The method comprises the following steps: acquiring a terahertz spectrum of a biochemical substance through a terahertz time-domain spectroscopy system (THz-TDS); labeling an absorption peak of the terahertz frequency domain spectrum by using Labelimg data labeling software, taking the labeled absorption peak as a terahertz spectrum data set, and constructing a YOLO biochemical substance terahertz spectrum recognition model based on a YOLO v5 model; and preprocessing a biochemical substance terahertz spectrum collected by the THz-TDS, and inputting the preprocessed biochemical substance terahertz spectrum into the trained YOLO biochemical substance terahertz spectrum recognition model for detection to obtain a target detection result. The terahertz spectrum is identified according to the terahertz absorption peak of the biochemical substance, the accuracy is high, the dependence on a large amount of training data is reduced, and the method can be widely applied to the field of terahertz spectrum identification.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Terahertz detection method

    CN108226089A

  • Terahertz time domain spectroscopy-based coal and rock discrimination method

    CN108279217A

  • Human body security check image detection method and system based on improved YOLOv5s

    CN114862837A

  • Method for extracting high-dimensional features from terahertz mixed spectrum for substance identification

    CN115855865A