Rapid detection method of Pleurotus eryngii mycelium based on near infrared spectral characteristics

Edible fungus is a large fungus with edible and medicinal value. Rapid detection of mycelium phenotypic characteristics is of great significance for edible fungus breeding and intelligent cultivation. Traditional method based on experienced observation easily led to make mistakes on distinguishing the growth stages, which impacted on the yield and quality of edible fungus. Therefore, in view of the lack of accurate and efficient detection technology during the growth stages of Pleurotus eryngii mycelium, a rapid detection method of Pleurotus eryngii mycelium at different growth stages is proposed based on the characteristics of near-infrared spectroscopy. First, the spectral data of mycelium of Pleurotus eryngii at six different growth stages were scanned. Second, the multivariate scattering correction method (MSC) was used to pre-process the raw spectral data, and then the competitive adaptive reweighted sampling algorithm (CARS) was adopted to detect the characteristic wave number of the effective variables for Pleurotus eryngii mycelium. In addition, the mathematical model between the mycelium of Pleurotus eryngii and the characteristic wave number of near-infrared spectrum was established by using feed forward neural network (BP). Finally, and the coding vector output by the network was used to detect to the growth stages. The results showed that the BP neural network structure of MSC-CARS-BP detection model was 86-85-85-95-6, and the accuracy of identifying different growth stages of Pleurotus eryngii mycelium was 99.67%. The research results could provide a new idea and technical support for the rapid detection of Pleurotus eryngii mycelium at different growth stages. (C) 2022 Elsevier B.V. All rights reserved.

Automated summary

This study proposes a rapid detection method for different growth stages of Pleurotus eryngii mycelium based on the characteristics of near-infrared spectroscopy. The method uses multivariate scattering correction and competitive adaptive reweighted sampling algorithms to establish a mathematical model between the mycelium and the characteristic wave number. The results show a 99.67% accuracy in identifying different growth stages of Pleurotus eryngii mycelium.