Raw egg freshness indexes mainly include sensory indicators, air cell height, Haugh unit, pH value, yolk index, specific gravity of eggs and total volatile basic nitrogen content. Traditional methods can test only part samples with low efficiency. This paper reviews the application of visible/near infrared spectroscopy, machine vision, electronic nose, dielectric properties and low-field nuclear magnetic resonance (LF NMR) for new non-destructive testing of egg freshness. We analyze advantages and disadvantages of each technology, and focus on the application of LF NMR technique in this field. Finally, existing problems and future directions are proposed to provide useful information for further research. Graphic abstract

Multispectral remote sensing data with abundant spectral information can be used to compute vegetation indices to improve the accuracy of Ginkgo biloba yield prediction. The limited spatial resolution of multispectral cameras restricts the detail capture over wide farmland, but super-resolution (SR) reconstruction methods can enhance image quality. However, most existing SR models have been trained on images processed from downsampled high-resolution (HR) images, making them less effective in reconstructing real low-resolution (LR) images. This study proposes a GAN-based super-resolution reconstruction method (RMSRGAN) for multispectral remote sensing images of Ginkgo biloba trees in real scenes. A U-Net-based network is employed instead of the traditional discriminator. Convolutional block attention modules (CBAMs) are incorporated into the Residual-in-Residual Dense Blocks (RRDBs) of the generator and the U-Net of the discriminator to preserve image details and texture features. An unmanned aerial vehicle (UAV) equipped with a multispectral camera was employed to capture field multispectral remote sensing images of Ginkgo biloba trees at different spatial resolutions. Four matching HR and LR datasets were created from these images to train RMSRGAN. The proposed model outperforms the traditional models by achieving superior results in both quantitative evaluation metrics (peak signal-to-noise ratio (PSNR) is 32.490, 31.085, 27.084, 26.819, and structural similarity index (SSIM) is 0.894, 0.881, 0.832, 0.818, respectively) and qualitative evaluation visualization. Furthermore, the efficiency of our proposed method was tested by generating individual vegetation indices (VIs) from images taken before and after reconstruction to predict the yield of Ginkgo biloba. The results show that the SR images exhibit better $R^2$ and $RMSE$ values than LR images. These findings show that RMSRGAN can improve the spatial resolution of real multispectral images, increasing the accuracy of Ginkgo biloba yield prediction and providing more effective and accurate data support for crop management.

The deep learning-based finger vein recognition method has achieved excellent recognition performance in local tasks. However, in distributed scenarios where the method is deployed on terminal devices, the recognition performance experiences a significant decline for new users who have not been included in the model’s training set. One feasible solution is to equip the model with the capability of continual learning. Nevertheless, due to the increasing concerns over data privacy issues and legal restrictions, it becomes challenging for the recognition system to access vein images from different terminal devices to continuously improve the recognition model. In this work, to mitigate the conflict between model learning and data privacy, we propose a finger vein recognition framework based on federated learning, called FVRFL, which gives the recognition model sustainable learning ability in a privacy-aware manner. Furthermore, to tackle the challenge of user diversity in finger vein recognition terminals, we propose a modular optimization strategy. This strategy optimizes personalized models for different terminals and designs target functions for distinct modules, catering to the high-performance requirements of terminal users for the recognition system. Extensive experiments on public datasets demonstrate that the proposed method not only empowers the recognition model with the ability of continual learning, improving the performance of the recognition system on terminal devices, but also effectively protects the privacy of user vein templates.