| Lycium barbarum is a pillar of Ningxia regional characteristics of agriculture,accounting for 33%of the total planting area of the country,the export of 90%of the country.Aiming at the current intelligent irrigation management of Lycium barbarum based on a single index and experience,under the background of the development of modern water-saving agriculture,this study combined artificial intelligence,Internet of things technology and irrigation technology to construct an intelligent,precise and efficient water saving irrigation scheme for Lycium barbarum in the arid region of central Ningxia.In order to promote the Lycium barbarum planting to achieve a new pattern of intelligent agriculture water-saving pattern.The main research work and conclusions are as follows:(1)During 2018 to 2020,the field irrigation experiments on high efficiency water-saving were conducted in the Lycium barbarum plantation in Runde,Tongxin County,an arid region of central Ningxia.Three irrigation levels were designed based on the reference crop evapotranspiration(ET0),which were 65%ET0,85%ET0,105%ET0,respectively.Fertilization was carried out according to the standard of Lycium barbarum plantation.The results showed that the growth,physiological indexes,yield and quality of Lycium barbarum treatment with 85%ET0 was the best.The maximum dry fruit yield of Lycium barbarum was 2456.22,2367.52 and 2592.40 kg/hm2,respectively.The WUE also reached the maximum in 85%ET0 treatment,and the value was 0.63,0.55 and 0.68 kg/m3,respectively.Additionally,the quality of Lycium barbarum fruit was also reached the excellent level.(2)According to the simulation analysis of 3 years field trial data,Jenson model was recommended simulating the water-yield relationship of Lycium barbarum at different growth periods,and an empirical model of water-yield relationship of Lycium barbarum was established.The comprehensive evaluation method of analytic hierarchy process(AHP)and entropy weight method was used to comprehensively evaluate the quality indexes of Lycium barbarum,and three indexes were selected to represent the quality of Lycium barbarum,including grain size,total sugar and 100-grain weight,and an empirical model of water-quality of Lycium barbarum was established.(3)Combined with the daily meteorological data of three typical meteorological stations in the arid region of central Ningxia from 2006 to 2018,the applicability and accuracy of the reference evapotranspiration prediction model based on the hybrid bi-directional long short-term memory network(Bi-LSTM),which took the maximum temperature,minimum temperature and sunshine duration as input,was established and verified in this region.At the same,the crop coefficient of Lycium barbarum was deduced according to the data of field experiment,then the prediction model of water demand of Lycium barbarum was established by using the method of single crop coefficient.Though the verification,the average error between the predicted and the measured for the whole reproductive period was only 11.53 mm,which was 2.69%higher than the measured value.It has reached the requirement of making accurate irrigation plan for the Lycium barbarum intelligent irrigation decision system.(4)According to different planting objectives,the irrigation optimization decision models of water-yield and water-yield-quality multi-objective planning for Lycium barbarum were established respectively.The second-judgment intelligent decision strategy of irrigation was designed for practical application.The results showed that the irrigation scheme based on the Lycium barbarum water-yield irrigation optimization decision model of the yield of dry fruit was 2714.30 kg/hm2,121.9 kg/hm2 higher than the 85%ET0 treatment,and the water saving was 1.36%.Compared with CK,the yield increased 290.68 kg/hm2 and the water saving was 23.56%.The irrigation scheme based on the Lycium barbarum water-yield-quality multi-objective irrigation optimization decision model,fruit quality was improved by 0.67%,the yield was only reduced by 69.60 kg/hm2 and water saving by 6.85%compared with 85%ET0 treatment.Compared with CK,the quality of Lycium barbarum improved by 9.58%,increased by 98.88 kg/hm2 and saved water by 27.81%,which achieved the purpose of intelligent irrigation management of Lycium barbarum with high efficiency and water saving.At the same time,according to the 3 years field experiment data,the Lycium barbarum yield prediction model based on Bayesian network was established.After independent analysis,it was found that the yield of Lycium barbarum could be accurately predicted according to the nitrogen application rate and the plant height,and on this basis,fertilizer usage management was carried out.(5)Lycium barbarum irrigation intelligent decision system was been developed used B/S structure,the lightweight MQTT Internet of Things communication control protocol and LNME(Linux,Niginx,Mysql,PHP).The functions of intelligent irrigation decision-making,statistical analysis,irrigation plan management,sensor equipment management,a knowledge base and other functions were realized by this system,which improved the practicability and scientificity of intelligent irrigation decision-making system of Lycium barbarum,and the purpose of efficient water saving and yield increase was achieved. |
PDF Full Text Request