Classification and recognition method of rice situation based on gray correlation degree

Abstract

Grain storage is a complex process, affected by factors such as mold, temperature, humidity, and moisture. The use of multiple sensors to detect changes in rice pile parameters has gained prominence as a means to ensure the accuracy and timeliness of grain condition monitoring. However, the current technology does not effectively utilize data. The assessment criteria primarily rely on grain temperature, and the analysis of grain condition is simplistic. Additionally, it fails to adequately integrate information on temperature, humidity, moisture, gas concentration, and other parameters of the grain pile to form a unified assessment result. To address the isolated and one-sided reaction of various parameters in the grain pile, this thesis conducts research on the storage characteristics of heating, condensation, and mold condition. It combines the information fusion of temperature, humidity, moisture, and CO₂ with normal grain conditions, constructs an assessment model based on the classification and identification of grain conditions under gray correlation, and achieves real-time dynamic assessment of the state of the grain pile. The experimental results show that the assessment model based on gray correlation can accurately discriminate between normal and mold conditions, but the accuracy in distinguishing heating and condensation still requires improvement. The overall recognition rate of the four types of grain conditions is 79%, which demonstrates the effectiveness of the model in identifying abnormal grain states.