Research Of Real-time Acquisition Method And Monitoring Model Establishment Of Aphis Gossypii Occurrence Quantity Information

The effective control of cotton aphids(Aphis gossypii)is based on timely and accurate information about pest populations.Currently,aphid populations are determined by visual counts.However,aphids are randomly distributed in fieldsand constantly moving by flight.Even when significant amounts of manpower and material resources are used,population estimates are not precise enough.The prevention and control of cotton aphids is hindered by inaccurate information about their population.Therefore,the objectives of this study were(i)to Test with the yellow boards at different heights and in different directions to determine the best conditions for monitoring aphid populations and(ii)to develop a device for determining aphid populations in drip-irrigated cottonusinga high definition camera,wireless transmission to a computer,and image recognition technology.The device can automatically change the yellow boards every 30 minutes,Solved the problem of collecting image background complex and aphids adhesion.Based on these results,models were established to estimate the population of different types of aphids.Based on the warning modelto notify farmers when the population reaches a critical number.Finally,A PC system was developed to control the device.The system can provide farmers with an early warning about increases in aphid populations and recommend effective treatment for their control.The main results are as follows:1.The key to monitoring cotton aphids is their attraction to yellow objectsCotton aphids are attracted to yellow objects.This objective of this field study was to determine the best conditions for monitoring aphid populations by testing the effectiveness of yellow board traps(i)at different positions in the field,(ii)at different heights,and(iii)in different directions.The study was conducted in a drip-irrigated cotton field in 2013 and 2014.The results showeda significant relationship between the number of wingedaphids and the total number of aphids.This means that the winged aphids can be a good indicator of the total cotton aphid population.The height and direction of the yellow boards also had significant effects on trapping efficiency.Two-way analysis of variance showed that height,direction,and height ×directioninteraction hadextremely significant effect on the trapping efficiency of the yellow boards.Therefore,the height above the canopy and the direction of the yellow board has great influence for monitoring aphid populations.The resultsdemonstrated that yellow boards that are positioned horizontally(i.e.,flat)should be 90 cm the top of the canopy.The bottoms of yellow boards that are positioned vertically(i.e.,upright)should be60 cm above the canopy.These two conditions are best for rapidly monitoring aphid populations in drip-irrigated cotton fields.2.Arapid acquisition method for determining cotton aphid populations in drip-irrigated cottonAfter determining the best monitoring conditions,the next step was to construct a rapid sampling device for determining aphid populations in drip-irrigated cotton.The device used high-definition images,wireless remote transmission technology,mechanical automation control and other modern information technology.The characteristics of the device are that it is to collect the image,simplify the background of the image,and solved the problem of rapid aphid accumulation on the yellow boards.Tests with the unit showed that 95.83% of the images were usable.The images were wirelessly transferred to a PC system with a99.4%success rate.The image transmission took an average of 191 s.The transmission rate was rapid enough so as to not affect the efficiency of the rapid monitoring and early warning system.Therefore,the monitoring device is reliable andit can also rapidly gather image information that can be used to determine winged aphid populations in drip-irrigated cotton.Recognition software was designed and developed based on image recognition technology.The software is simple,easy to use,scalable,and unified.This software system implements the cotton aphid alarm parameters(e.g.rate of automatic data collection,statistical analysis,and other functions).The results of the system were compared with visual counts.The relative error of the system was only 2.67%.This indicated that the automated system can completely meet the needs of an alarm system for making decisions about treating a growing aphid population.3.Real-time estimates of cotton aphid populations and validation of the modelThe device was tested in field experiments in 2015.Visual counts were also done at the same time.Correlation analysis was conducted between the visual counts and those obtained from the device.Based on these results,models were established to estimate the population of different types of aphids on 100 cotton plants.The models were validated in 2016.The validation tests showed that the model's prediction accuracy of winged aphids in the field was low.The model's best predictions were for total aphid population.Total aphid populationson 100 plants can be estimated based on the number of winged aphids measured by the device.The estimated value of the root mean square error(RMSE)wasonly 2180.Therefore,the forecast model for the total aphid population in the field(Y = 0.0362 x2-12.642 x + 16470)has good feasibility.Field surveys were conducted to classify the degree of damage due to cotton aphids and then compared with the estimates of the rapid monitoring device.The results showed that the prediction level was completely consistent,with accuracy up to 100%.Therefore,this study showed that the model can completely satisfy the requirements of an early warning and forecast system forcotton aphid populations.Combined with meteorological data from a local source,the system can provide reliable data for preventing and controllingcotton aphid in open fields.4.An early warning and decision-making system for cotton aphids in drip-irrigatedcottonThe automated systemdescribed above combinesdynamic information acquisition,image recognition technology,and a rapid monitoring and early warning model.Programming language and database technology were used to develop an integrated information technology system.The decision-making system consists of five functional modules:(i)information display,(ii)system management,(iii)information management,(iv)aphis gossypii early warning and decision-making,and(v)information query and benefit evaluation.By referencing estimates of aphid populationand meteorological information,the system provides early warning information and recommends the corresponding prevention measures(e.g.,optimum control period and pesticide use).In this way,late treatment is avoided.The system can effectively control aphids over a large area and improve the accuracy and timeliness of treatment.This reduces pesticideamounts and reduces costs.