| The Yellow River delta is located in the overlapping zone of yellow and blue national strategies,with abundant reserve land resources and superior geographical location,which has important natural ecological value and social and economic value.However,due to its unique sedimentary environment,climatic conditions,soil parent materials and the existence of production methods such as excessive exploitation and processing of petroleum,the ecological environment in this region is very fragile,mainly manifested in:the primary salinized soil is widely distributed,and the secondary salinization is serious,the ecosystem structure is single and the resistance is poor,low resource utilization and weak agricultural productivity.Therefore,in order to effectively protect the ecological environment of the Yellow River delta region,it is an urgent need for the development of the Yellow River delta region to carry out real-time monitoring of the main ecological parameters of the regional surface.The content of soil surface water and salt are important ecological parameters that restrict the development of ecological environment in the Yellow River delta.The ecological protection in the Yellow River delta should not only focus on the monitoring and regulation of soil surface water and salt content,but also take the healthy growth of vegetation as the ultimate goal.The traditional method of monitoring ecological parameters is field survey and field measurement.Although this method has high accuracy,it is gradually replaced by satellite remote sensing.Satellite remote sensing monitoring has become the mainstream due to its advantages of nondestructive measurement,high efficiency and so on.Subsequently,Unmanned Aerial Vehicles(UAV)gradually developed and entered the civil field,becoming an important means of ecological environment monitoring.Compared with traditional methods,UAV remote sensing can obtain high-throughput image data in a short time,saving a lot of time,manpower and economic costs.And compared with satellite remote sensing,UAV remote sensing has the characteristics of high temporal and spatial resolution,low cost,high timeliness,and repeatability,etc.At the same time,it can effectively eliminate the influence of clouds and collect image data even on cloudy days.Therefore,in this paper,the representative of the Yellow River delta area—Hekou district(core test area,a total of setting a sampling point)and Kenli district(verification area,deploy 69 sampling points)are for ground investigation,four main surface ecological parameters,namely soil salt content,soil water content,chlorophyll content(SPAD)and vegetation coverage are selected,and the obtained investigate data,UAV multispectral data(including green,red,red and near infrared band)and Sentinel-2A satellite multispectral data as data sources By fully combining the advantages of the three,the multi-scale inversion of each ecological parameter is realized with high precision,high temporal and spatial resolution.Firstly,the correlation analysis and band grouping method were used to analyze the ecological parameter data and UAV image data obtained from field measurement,and the characteristic band,sensitive vegetation index and characteristic spectral parameters were screened out.Secondly,a variety of regression methods were used to construct an estimation model and optimize the model by taking the characteristic band,the sensitive vegetation index and the characteristic spectral parameters as independent variables and the measured values of each ecological parameter as independent variables.Then,the relationship between the reflectivity of UAV image and satellite remote sensing image is analyzed,the normalized reflectivity coefficient of satellite image is obtained by using the ratio mean method,and the normalized reflectivity of satellite image is processed,the time and space universality of the estimated model is verified by using the method of model precision comparison and verification area verification.Finally,the band calculation method is used to bring the reflectance of each band of the UAV image and the normalized satellite image into the optimal estimation model of each ecological parameter,and the inversion results of each ecological parameter under the scale of field(UAV image)and research area(satellite image)are obtained.The results show that the correlation between reflectance and ecological parameters in uav image is better than that in satellite image.The modeling accuracy of the best estimation model of soil salt content was R~2=0.743,RMSE=3.05,MAE=2.13,and the verification accuracy was R~2=0.809,RMSE=2.04,MAE=1.51.The model was constructed based on the multivariate linear regression method,and the characteristic band and characteristic spectral parameters of UAV images were independent variables.The modeling accuracy of SPAD's best estimation model was R~2=0.926,RMSE=0.63,MAE=0.92,and the verification accuracy was R~2=0.934,RMSE=0.78,MAE=0.87.It was constructed based on the multivariate linear regression method and the characteristic spectral parameters of UAV images were independent variables.The modeling accuracy of the best estimation model of vegetation coverage is R~2=0.799,RMSE=0.03,MAE=0.02,and the verification accuracy is R~2=0.769,RMSE=0.27,MAE=0.22.The model is constructed by Sequential Minimal Optimization(SMO)method,and the characteristic band and characteristic spectral parameters of UAV images are independent variables.The three estimation models are proved to have high universality in time and space.The normalized satellite image inversion results are satisfactory.The multi-scale inversion results are consistent with the reality and effectively reflect the spatial distribution characteristics of the ecological parameters in the core test area and the research area.But it is unfortunate that the bands included in this study are not suitable for inversion of soil water content in either UAV image or satellite image.Combining the advantages of satellite,UAV and field measurement methods,this paper provides an effective method for multi-scale inversion of major ecological parameters of the surface,which is of great value for real-time,rapid and accurate monitoring of important ecological parameters in the Yellow River delta region. |
PDF Full Text Request