Background: Oilseed rape, as one of the most important oil crops, is an important source of vegetable oil and protein for mankind. As a non-essential element for plant growth, heavy metal cadmium (Cd) is easily absorbed by plants. Cd will inhibit the photosynthesis of plants, destroy the cell structure, slow the growth of plants, and affect their development and yield. It is necessary to develop a method based on visible near-infrared (NIR) hyperspectral imaging (HSI) technology to quickly and nondestructively determine the Cd content in rape leaves.Results: Two-layer estimation models were established by combining visible-NIR HSI with ensemble learning methods (stacking and blending). One layer used support vector regression, extreme learning machine, decision tree, and random forest (RF) as basic learners, and the other layer used support vector regression or RF as a meta learner. Different models were used to analyze the spectra of rape treated with five Cd concentrations to obtain the best prediction method. The results showed that the best model to predict Cd content was the stacking ensemble model with RF as the meta learner, with coefficient of determination for prediction of 0.9815 and root-mean-square error for prediction of 5.8969 mg kg −1 . A pseudo-color image was developed using this stacking model to visualize the content and distribution of Cd.
Conclusion:The combination of visible-NIR HSI technology and the stacking ensemble learning method is a feasible method to detect the Cd content in rape leaves, which has the potential of being rapid and nondestructive.
In order to the quickly and nondestructively detect whether starch is adulterated in minced chicken meat, a novel method combining hyperspectral imaging (HSI) technique with transfer learning was proposed in this study. First of all, hyperspectral images of minced chicken meat with different mass fractions of starch were collected and spectral information of the samples in the range of 400.89-1000.19 nm was extracted. Then, the hyperspectral data was preprocessed via continuous wavelet transform (CWT), which transformed the pixel-level hyperspectral data into twodimensional spectrograms. Furthermore, classification model for identifying starch in minced chicken meat was constructed based on the GoogLeNet network pretrained on ImageNet data set. Finally, the support vector machine (SVM) model and the convolutional neural network (CNN) model without transfer learning were established for comparison. The results indicated that the model based on GoogLeNet network had a higher classification accuracy, up to 98.6%. Therefore, this study demonstrates the feasibility of the detection of starch in minced chicken meat based on HSI technique and transfer learning.Practical applications: The identification of starch in the minced chicken meat is of particular significance for maintaining the market order and safeguarding the human health. An innovative method based on hyperspectral imaging technique and transfer learning to identify whether the minced chicken meat mixed with starch was proposed in this study. The method achieved quickly, nondestructively and accurately detection of starch in minced chicken meat.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.