论文部分内容阅读
目的应用半透射高光谱成像技术结合支持向量机(support vector machine,SVM)模型实现马铃薯内外部缺陷多指标同时检测。方法采集310个马铃薯样本半透射高光谱图像,并分别采用标准正态变量变换(standard normalized variate,SNV)、归一化(normalize)和平滑处理(smoothing)对光谱信息进行预处理。进一步采用竞争性自适应重加权算法结合无信息变量消除法(competitive adaptive reweighed sampling algorithm,uninformative variable elimination,CARS-UVE)进行特征波长选择,提高模型识别率。结果原始光谱信息经归一化预处理和竞争性自适应重加权算法结合无信息变量消除法(CARS-UVE)降维后所建的支持向量机(SVM)模型识别结果最优,该方法对合格、绿皮和黑心马铃薯样本预测结果分别为90.7%、88.9%、95.7%,混合识别率为91.3%。结论采用半透射高光谱成像技术结合CARS-UVE方法所建SVM模型能够实现马铃薯内外部缺陷多指标同时检测。
Objective To detect simultaneous multi-index of internal and external defects of potato by semi-transmissive hyperspectral imaging combined with support vector machine (SVM) model. Methods Semi-transmission hyperspectral images of 310 potato samples were collected and normalized and smoothed using standard normalized variate (SNV), respectively, to preprocess the spectral information. The competitive adaptive reweighted algorithm and the uninformative variable elimination (CARS-UVE) are further used to select the characteristic wavelength and improve the model recognition rate. Results Support vector machine (SVM) model based on original spectral information was optimized by normalized preprocessing and competitive adaptive weighted algorithm combined with CARS-UVE dimensionality reduction. The predicted results of qualified, green and black-hearted potato samples were 90.7%, 88.9% and 95.7% respectively, and the mixed recognition rate was 91.3%. Conclusion Simultaneous detection of multiple internal and external defects in potato by semi-transmissive hyperspectral imaging coupled with the SVM model developed by CARS-UVE method can be achieved.