Study on the dimension correction of juicy peach based on diffuse transmittance online detection

Abstract

At present, visible/near infrared (VIS/NIR) spectroscopy has been widely used in fruit quality detection, but the different dimensions of fruit will affect the accuracy of quality detection, especially the online detection. At this time, it is necessary to study a dimension correction method to improve the accuracy of spectral detection. Therefore, in this paper, VIS/NIR spectroscopy technology was used to collect the spectra and soluble solid content (SSC) of three fruit types of peach(small, medium and large fruit) under different integration time. The analysis of their spectral attenuation revealed that the spectral energy of medium and large fruit, all attenuated under short integration time, and their complete spectrum could be collected under long integration time, indicating that adjusting the integration time was feasible. After that, partial least square method (PLS) was used to develop the combination model with the spectral data of different dimensions of peach under different integration time, and the optimal model was obtained. When the integration time of small fruit, medium fruit and large fruit was 80 ms, 100 ms and 120 ms respectively, the prediction correlation coefficient (Rp) of the model was 0.846. Then multiplicative scatter correction (MSC) was used to pretreat the optimal model, and the Rp of the model was 0.857. The results showed that using different integration time could reduce the influence of dimension difference on the online quality detection of peach, realize dimension correction, and more accurately detect the quality of peach. In addition, it also showed that the dimension correction method can improve the accuracy of spectral detection, and has certain application value for spectral detection.