Light Response Characteristics And Ear Morphology Monitoring Of Drip Irrigation Maize Based On Mobile Phone Photographs

Using mobile phone cameras to monitor crop growth is more portable and promotional than SLR cameras.In this study,six key nitrogen application rates(0,90,180,270,360,450 kg/hm2,denoted as NO,N1,N2,N3,N4,N5)were used to obtain canopy images of key growth stages of maize.And measure the agronomic parameters simultaneously.The image parameters were extracted and screened,and the maize light response parameters and spike morphology parameter inversion models based on image parameters were established and tested,which provided a reference for the precise cultivation of drip irrigation maize in this area.The main conclusions are as follows:(1)Dynamic analysis of RGB model of maize image based on the number of days after emergence and effective accumulated temperature in the fertility process.The two have the best fit to the red value(R),which is consistent with the rational function y=(b+cx)/(a+x),R2≥0.866,and the green(G)and blue(B)values.The logarithmic function y=a-bln(x+c)relationship R2 is≥0.692 and 0.734,respectively.(2)Selecting the optimal light response curve model of maize large bell mouth period.The fitting accuracy of the four models for N3,N4,and N5 light response curves is higher than N0,N1,and N2.The modified model overcomes the other 3 models that cannot fit the light suppression phenomenon with high precision(R2≥0.996,RMSE=0.84μmol·m-2·s-1,MAE=0.682)can be used as the optimal light response curve model for maize in this area.(3)Inversion model of maize light response characteristic parameters based on mobile phone photos.The correlation coefficient between image parameters and optical response parameters is CC-bgr-ExG,which is significant and greater than 0.540.The CC-based maize light response parameter model is established and tested.The parameter modeling and verification set R2>0.830,RMSE<5.989 μmol·m-2·s-1,nRMSE<10.000%.(4)Inversion model of maize ear morphology parameters based on mobile phone photos.The morphological parameters of each ear showed an S-curve pattern with the grouting process and reached the maximum under N4.The correlation between CC and ear morphological parameters was the best,and the correlation coefficient was greater than 0.610.A CC-based pan morphology model was established and tested.Each parameter modeling and validation set R2>0.506,RMSE<83.964ml,nRMSE<34.925%.The results showed that the monitoring of maize light response parameters and ear shape parameters can be achieved by using mobile phone camera.