Traditional methods for detecting seed germination rates often involve lengthy experiments that result in damaged seeds. This study selected the Zheng Dan-958 maize variety to predict germination rates using multi-source information fusion and a random forest (RF) algorithm. Images of the seeds and internal cracks were captured with a digital camera. In contrast, the dielectric constant of the seeds was measured using a flat capacitor and converted into voltage readings. Features such as color, shape, texture, crack count, and normalized voltage were used to form feature vectors. Various prediction algorithms, including random forest (RF), radial basis function (RBF), neural networks (NNs), support vector machine (SVM), and extreme learning machine (ELM), were developed and tested against standard germination experiments. The RF model stood out, with a training time of 5.18 s and the highest accuracy of 92.88%, along with a mean absolute error (MAE) of 0.913 and a root mean square error (RMSE) of 1.163. The study concluded that the RF model, combined with multi-source information fusion, offers a feasible and nondestructive method for quickly and accurately predicting maize seed germination rates.

Insect electrical penetration graph (EPG) instrument is a powerful tool to study the feeding behavior of piercing-sucking insect and insect transmission virus mechanism. However, the classification of EPG waveforms has been carried out manually. It is urgent to identify EPG waveforms automatically to improve the analysis efficiency. Machine learning methods are widely used in classifying human bioelectrical signals, but rarely in EPG signals. In this study, the kernel extreme learning machine (ELM) was applied to the classification of EPG waveforms. Morlet wavelet function was used as the kernel function of the ELM, and the number of neurons was selected adaptively by incremental algorithm to solve the classification problem of EPG waveforms quickly and accurately. In the experiment, the 6-dimensional feature vector was used to enter the wavelet kernel extreme learning machine (WKELM) which was composed of low frequency wavelet energy in the 2nd and 3rd layers, fractal box dimension, Hurst exponent and spectral centroid in the first two layers of HHT. The average classification accuracy was 94.47%, which was 3.04% higher than the previous study. The experimental results showed that the proposed classification method of EPG waveforms based on WKELM has high identification performance, which laid a theoretical foundation for the research and development of EPG waveforms automatic identification and analysis system.

In this paper, freshness monitoring model on small larimichthys polyactis based on multiple sensor array system (MSAS) combined with total volatile basic nitrogen (TVB-N) examination was studied. Larimichthys polyactis was obtained and prepared as detecting samples. MSAS was developed based on volatile gas content of the fish samples. Meanwhile, TVB-N was examined to provide quality reference for MSAS measurement. MSAS responding data were processed by principal component analysis (PCA) and non-linear signal analysis model. Larimichthys polyactis freshness deciding model was developed by non-linear fitting between signal processing features and TVB-N parameters. Results indicated that the method studied in this work presented good freshness deciding accuracy for larimichthys polyactis. This method also presented some advantages including rapid analysis, easy operation, and low cost. It is promising in aquatic products quality raped determination.