Discrimination of Infected Silkworm Chrysalises using Near-Infrared Spectroscopy Combined with Multivariate Analysis during the Cultivation of Cordyceps militaris

Abstract

The objective of this study was to confirm whether near-infrared spectroscopy could be used to discriminate the infected silkworm chrysalises. A total of 105 silkworm chrysalises — 65 infected and 40 uninfected — were collected at Beijing Shoucheng Agricultural Development Co., Ltd. Near-infrared spectra were acquired at the head, chest, abdomen, and posterior belly of each silkworm chrysalis (both uninfected and infected). Three spectral pre-processing methods and four discrimination models were used to identify the uninfected and infected silkworm chrysalises. Results indicated that the PLS–DA model based on the spectra processed by multiplicative scatter correction (MSC) had the best discrimination performance (the prediction accuracy of calibration set and prediction set were 100 and 97.5%, respectively), and the head portion was the best position for the discrimination of uninfected and infected silkworm chrysalises. The overall conclusion was that the uninfected and infected silkworm chrysalises could be successfully identified by using near-infrared spectroscopy technology in the cultivation of Cordyceps militaris.