研究目的
To improve the classification performance of laser-induced breakdown spectroscopy (LIBS) technology by proposing an image features assisted line selection (IFALS) method based on spectral morphology and the characteristics of Harris corners.
研究成果
The IFALS method significantly improves the classification accuracy and efficiency of LIBS technology compared to the conventional MLS method. The classification accuracy was increased from 94.38% to 98.54%, and the time required for the classification process was decreased from 2768.38 seconds to 4.36 seconds. The method also shows good generalization ability and faster convergence rate compared to existing automatic line selection methods.
研究不足
The study does not mention the limitations of the IFALS method.
1：Experimental Design and Method Selection:
The IFALS method was proposed based on spectral morphology and Harris corners characteristics. A classification experiment for 24 metamorphic rock samples was conducted using the LDA algorithm.
2：Sample Selection and Data Sources:
24 kinds of natural metamorphic rocks were used in the experiment. Four points were selected on each sample for laser ablation, 25 spectra were collected from each point, and each spectrum was averaged from 3 spectra.
3：List of Experimental Equipment and Materials:
A compact diode-pumped laser and a compact spectrometer (Avantes Mini-MK-II) were used.
4：Experimental Procedures and Operational Workflow:
The laser beam was focused on rock samples to generate plasma. The emission was collected and transmitted to the spectrometer. The IFALS method was applied to select analytical lines, and the LDA algorithm was used for classification.
5：Data Analysis Methods:
The classification accuracy and time required for the classification process were compared between the conventional MLS-LDA method and the proposed IFALS-LDA method.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
