Corn Seeds Clarity Testing Method And Device Study Based On Ultrasonic

In our country,the existing methods of corn seeds clarity processing and the precision filtering for particles with moth-eaten-holes and mildew infestation is not enough.Therefore,the seed-germination-rate precision processing still cannot meet the requirements of single grain sowing and if the mildew particles are introduced into the food chain,they can cause serious damage to the health of people and animals.In order to address these problems and summarize the research results from around the world,this paper puts forward the research of corn seed clarity detection methods and devices,based on air-coupled ultrasonic nondestructive testing technology,signal processing method,computer technology and electronic technology.The main research contents and achievements include:(1)Study the principle and method of corn seed clarity,based on ultrasonic nondestructive testing technology.First,select the control group of corn seeds that have and not been irradiate by the ultrasonic method in order to do the comparative germinate test to verified that use ultrasonic technology to detect corn seeds is security.Then,analyze the principles of the moth-damaged and mildew particle detection that by the same technology.Subsequently,the mechanical characteristics of corn particles were analyzed,and according to its unique characteristics,the longitudinal wave velocity were calculated.Then,according to the velocity and accuracy,the frequency of excitation signal have be set.Finally,based on the needs of detection precision and requirements of the higher SNR signal,the process have integrated and optimized the ultrasonic signal acquisition device.(2)In this study,a novel method based on air-coupled ultrasonic technology was proposed to detect moth-damaged corn seeds.First,denoising and feature extraction of the corn seeds ultrasonic signal that acquired at room temperature.Then,establish the KNN,DT,SIMCA and LDA four recognition models by using the data of single and double-sided corn grains.Results show that the seed's azimuth has no affect on the detection and the correct recognition rate of all models for the two classes of corn particles is higher than 85.00%.Among them,the KNN model had the best performance in regard to it's high recognition rate and stability.The recognition rate for all the data was higher than 97.00%.The research results show that air-coupled ultrasonic technology is feasible to detect moth-damaged corn particles.(3)Propose a new method for corn mildew particle detection that is based on air-coupled ultrasonic technology.First,to do the Hilbert Huang transform for both the mildew and intact seeds signal data,extract the frequency between 300-498 hz Hilbert marginal spectrum as the classification characteristics of both kinds of seeds,then established the BPNN,SVM,SIMCA and KNN four kinds of classification models,and the model performance was tested.Results showed that the best recognition model for mildew and intact grains is BPNN,with a recognition rate of more than 80.00%,especially when the number of hidden layer neurons is 6,the recognition rate was able to reach 85.71% and 91.43%,respectively;The SVM model takes the second place,with a recognition rate for the two kinds of seed of 80.00% and 85.71%,respectively.These results show that use the the air-coupled ultrasonic technology to detect corn mildew particles is feasible.(4)Developed the corn seed clarity detection device.According to the requirements of the corn seeds testing actual,the air-coupled ultrasonic signal acquisition device and testing software system have been integrated.Design and carry through the accuracy,inclusive,stability and operating speed test experiment for the testing system,which showed that the recognition rate of the system for every kinds and types corn seed particles over 80% all.With a stable recognition rate,the system can reach a certain accuracy,tolerance and stability,allowing the detection device to meet the actual demand of corn seed clarity detection.