Research On Plant Water Stress Evaluation Based On Plant Electric Signal

Plant electrical signals are physiological signals that respond to changes in the outside world.The use of plant electrical signals as an effective indicator of the plant growth status diagnosis mechanism is a new theory for studying the relationship between plant growth environmental factors and their growth status.The new method has the characteristics of sensitive and accurate response,so it can realize the automation and intelligence of agricultural planting by studying the changing law of plant electrical signals.This paper is cultivated by Shanghai Qing as research objects,and controls environment variables as soil water content,collect external form data of Shanghai green growth under different soil water content,and combines plants to synthesize electricity signals in different water stress Evaluate the growth of plants under different soil water content.By studying the growth state of plants,the degree of water stress is verified by plants,and the optimal water demand gradient for plant growth can be automatically predicted through comprehensive indicators,which provides effective reference value for subsequent intelligent agricultural irrigation.The BL-420 biological function experimental device collects electrical signal data,and uses wavelet threshold soft and hard threshold and wavelet packet analysis to denoise the noisy signal of small rapeseed.With the change of soil water content,the collected electric signal waveforms of rapeseed are also different,but the amplitudes are all at the ?V level,and the frequency is mainly concentrated below 3Hz.From the time domain analysis mean,root mean square,standard deviation,etc.from the frequency domain analysis edge frequency(SEF),gravity frequency(SCF)power spectrum entropy(PSE)with soil water content,found in the soil When the water content is increased,SEF,SCF,and PSE have an elevated trend,but in the case of more drought,the frequency domain feature value is low.From the perspective of wavelet packet decomposition,the change of electrical signal energy in the moisture is used as one of the characteristic values.The electrical signal characteristic value is generated by the desired feature vectors,using external form data and the water gradient of Shanghai Green growth derived from the external form data and the water gradient,as an evaluation indicator of the classification effect of the electrical signal.Finally,the support vector machine is used to classify the moisture stress state of plant growth to the plant growth and the relationship between plant growth state and plant electrical signals.The results show that the level of water stress in SVM-based plants is90.83%,and the mean square error MSE is 0.175;the accuracy of the moisture stress based on the PSO-SVM plants is 95.4167%,and the mean square error is 0.1646.The results of classification experiments show that electrical signal analysis can be used to classify plants under water stress and plant growth status,and the SVM classification model optimized by particle swarms is more accurate,which is for the realization of automated agricultural planting through plant electrical signals.Lay the foundation for observation and other aspects.