Effect of spectrum measurement position on detection of Klason lignin content of snow pears by a portable <scp>NIR</scp> spectrometer

Snow pears are an important and widespread agricultural product that can relieve respiratory symptoms, constipation, alcoholism. Lignin content (LC) has a direct negative role on the fruit texture taste of snow pears. Here, we studied effect near-infrared (NIR) spectroscopy determination LC in due to position at which spectral measurements were obtained. NIR diffuse reflection spectra collected from nine measurement positions each sample by portable spectrometer. Partial least squares regression (PLSR) was used develop spectrum compensation models for three local models, average model, global model. The results indicated prediction accuracy affected position. Compared with model had good results. Next, synergy interval partial squares, bootstrapping soft shrinkage, competitive adaptive reweighted sampling, genetic algorithm, improved variable stability frequency analysis algorithm (VSFAA) method select most effective variables build PLSR calibration established using 10 selected VSFAA reduced influence variation achieved more promising results, correlation coefficient 0.842 0.824, respectively. root mean square error cross-validation 0.736 0.694, overall showed based sensitivity predicting combined selection improve provide robust combining (three stem-calyx longitudes intersected latitudes (stem, equator, calyx)) produced wavelengths (SNV-VSFAA-PLSR)-average reducing prediction, helpful offsetting equatorial alone.