Hyperspectral traditional Chinese medicinal material identification method based on adaptive random block convolution kernel network

• Main Authors: LIU MIN, SU XUESAN, LIU CAIPING, MAO JIANXU, ZHANG HUI, ZHU QING, YIN ATING, CHEN YURONG, ZHAO BINGRUI, WANG YAONAN, ZENG KAI, LI YAPING
• Format: Patent
• Language: Chinese; English
• Published: 28.01.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses a hyperspectral traditional Chinese medicinal material identification method based on an adaptive random block convolution kernel network, and the method comprises the steps: obtaining an optimal wave band subset of a hyperspectral image of a traditional Chinese medicinal material based on an optimal clustering frame, and effectively selecting an optimal characteristic wave band from the optimal wave band subset through employing a cluster sorting method; using a random projection method to take a random block extracted from the hyperspectral image of the traditional Chinese medicinal material as a convolution kernel; modifying a convolution kernel by using a pixel adaptive method, and performing feature extraction based on a traditional Chinese medicinal material feature wave band image; then, extracting features of the traditional Chinese medicinal materials by using a hierarchical network, and constructing a traditional Chinese medicinal material hyperspectral training set and a tra