Research On Effectiveness Evaluation And Parameter Sensitivity Analysis For The High Resolution Optical Remote Sensing Satellite

High resolution optical remote sensing satellite can acquire image data of meter-level or even sub-meter-level spatial resolution,which is widely used in the field of space information application.Before the satellite is launched into orbit and operating in orbit,a lot of assessment and prediction work will be carried out to ensure the normal operation of the satellite.However,traditional effectiveness evaluation process needs a lot of time and labor cost,and it is also unable to analyze the sensitivity of effectiveness index.Therefore,it is of great significance to mine efficiency prediction model from historical evaluation instead of serializable evaluation process.Meanwhile,the sensitivity analysis method of parameters is introduced to quantify the sensitivity of different indicators.Therefore,the main contents of this paper are as follows:According to the influence factors of the earth observation ability for the high resolution optical remote sensing satellite,the paper designs four layers of efficiency index system,including the layer of comprehensive index,the layer of ability index,the layer of application domain index and the layer of performance index.Then,this paper introduces the meanings of different efficiency indexes and their solutions of computation from two aspects of the inherent observation ability and the task observation ability of the satellite system,and puts forward the standardized method of index preprocessing.In view of the spatiotemporal coverage effectiveness index of the observation capability for the task,a calculation method based on grid division and vector intersection is proposed.In view of the limitations of the subjective and objective weighting model,combined with the improved AHP,the weighting method of entropy and the weighting method of principal component analysis,a combined weighting method based on the least square of effectiveness index is proposed.Experiments show that this method can effectively integrate the weight of different weighting models,thus ensuring the integrity of weight information.Based on the traditional grey correlation analysis method and radar map method,two improved evaluation methods are proposed,namely,the weighted comprehensive grey correlation evaluation method based on dynamic resolution coefficient(RGRA)and the improved radar map evaluation method based on Fuzzy Comprehensive Evaluation(FCE-RRME).RGRA can overcome the limitation of taking the resolution coefficient as the empirical value and solve the problem of inconsistent evaluation reference standards.FCE-RRME avoids the difference of evaluation results caused by the node sequence of the traditional method.After fuzzy processing,it can effectively solve the problem of small samples and gray evaluation.The reliability of the methods is verified by experiments.Aiming at the characteristics of tedious and time-consuming effectiveness evaluation process,this paper designs an effectiveness prediction model based on ensemble learning methods of Bagging and Boosting.Base on the data set constructed through latin hypercube parameter sampling,a comparative experiment is carried out with single prediction models.The experimental results show that the prediction error of Bagging-SVR model is the smallest,with the accuracy index of RMSE and MAE less than 0.005,MAPE lower than 0.9%and R~2 higher than 0.99,which shows that the effectiveness prediction model based on ensemble learning can effectively improve the prediction accuracy and better fit the evaluation results,thus injecting new research ideas into the field of effectiveness evaluation for remote sensing satellite.Aiming at the limitation that traditional effectiveness evaluation can't analyze the sensitivity of effectiveness index quantitatively,this paper applies the method of parameter sensitivity analysis to effectiveness evaluation,and designs a sensitivity analysis method of effectiveness index based on Sobol index method.The experimental results show that the sensitivity of spatial effectiveness index and spatiotemporal coverage effectiveness index is relatively high,which verifies the validity of this method,and provides certain reference significance for the sensitivity analysis of effectiveness evaluation for remote sensing satellite.