| In the face of the increasingly urgent demand for information monitoring of precision agriculture in mountainous areas,traditional agricultural remote sensing survey means are unable to meet the demand for obtaining refined agricultural information.Meanwhile,the complex surface environment and climate environment in karst mountainous areas bring great challenges to traditional agricultural remote sensing monitoring.With the development of microwave remote sensing technology,Synthetic Aperture Radar(SAR)brings a new opportunity for agricultural remote sensing monitoring in cloudy and rainy areas.However,the complex imaging mechanism leads to the low signal--noise ratio of SAR images,which has become a common problem in agricultural remote sensing monitoring by using SAR.Facing the three problems of agricultural remote sensing in karst mountainous areas,the multi-source remote sensing spatio-temporal cooperative method based on the theory of remote sensing information Tupu will become an effective scheme for agricultural remote sensing monitoring in complex mountainous areas.Embodied in the use of high spatial resolution to establish a standard set of space optical remote sensing image "Tu" benchmarking framework,using SAR remote sensing provides high return period,as well as the target feature space geometry information to establish a set of contains time-varying non-visual texture sequence "Pu" information,combined with uav remote sensing provides accurate sample infor mation,give full play to the respective advantages of multi-source remote sensing data.Time series based on geo-parcels "Tu" constraint "Pu" of multi-source remote sensing space-time cooperative ideas,this article aims to provide a set of oriented farmland geo-parcel scale planting characteristic information acquisition and recognition methods,including land parcel scale crop response relationship of SAR texture and phenological characteristics extraction,the types of crops and land state recognition.In order to provide scientific support for agricultural precision monitoring in karst mountainous areas.In terms of theoretical research,the idea of multi-source remote sensing spatio-temporal coordination using the farmland geo-parcel "map" to constrain the time series "spectrum" is the theoretical and methodological guidance of this study,and its feasibility plays an important role in supporting the subsequent research.In this paper,the spatial differentiation theory and geographic detector method are used to demonstrate that under the framework of standard spatial "Tu",farmland geo-parcel scale constraint can improve the expressive ability of SAR images,and the expressive ability of SAR target decomposition parameters containing spatial geometry information of ground object is significantly better than that of backscattering parameters.At the same time,the expression ability of temporal "Pu" information was greatly improved based on geo-parcel constraints.It indicates that supported by the "Tupu" theory of remote sensing information,the scheme of multi-source remote sensing spatial and temporal coordination can better solve the problem of agricultural remote sensing precision monitoring in the complex imaging environment of karst mountainous areas.This conclusion provides theoretical and method support for crop type discrimination and farmland status information extraction based on geo-parcel scale.In terms of data mining,in order to solve the time-series characteristic response problem of crop growth process at geo-parcel scale,high time resolution SAR remote sensing images were used to obtain high frequency time-series observation data of cultivated land,and the non-visual texture time series based on the spatial relationship of image pixels were constructed.Using multiple parameters,texture,spatial direction form multi-dimensional characteristic dataset,by machine learning method selection advantage parameters combination,further analyzes the internal space of farmland,crops caused by canopy structure change and the synergy of multi-source remote sensing data of land parcel scale response characteristics of the SAR texture.The results showed that the SAR texture reflected the change rule of bare soil-bare soil and vegetation-vegetation canopy of cultivated land with the growth process of crops from germination to maturity.The response characteristics and response time points of SAR texture are different among different crops,indicating that the temporal characteristics of SAR texture can reflect the growth of cultivated crops well.In addition,the response law of SAR texture in crop growth process and the recognition of corresponding key time points can provide method support for crop type discrimination,phenological feature recognition and cultivated land state information extraction based on farmland geo-parcel scale.In terms of application exploration,in order to extract farmland information at geo-parcel scale,based on the temporal features of SAR texture during crop growth process,a method of cultivation type discrimination,phenological information extraction and farmland status information identification was further developed.Combined with the precise cultivated land boundary provided by high spatial resolution optical remote sensing images,the land parcel scale SAR texture time series was used to explore the change response mechanism of different crop planting,growth,harvest and other key time points.It was found that the response mechanism of "U-shaped" or "W-shaped" was formed by the homogeneous state of cultivated land texture during crop growth and the disordered state of cultivated land texture during tilling/fallow,and the response mechanism of "U-shaped" or "W-shaped" was formed by the higher value during tilling/fallow and the lower value during growth,and the SAR texture time series curve was reconstructed.Based on the characteristics of curve change and time points,crop planting types were identified,crop phenology characteristics were extracted,cultivated land status was identified at plot scale,cultivated land planting system was speculated,and farmland "non-agricultural" and "non-food" information was extracted and analyzed.This research in remote sensing information "Tupu" theory,based on the theory of space-time cooperative thinking method in karst mountainous area surface complex,climate is complex and the SAR image imaging mechanism of complex three agriculture under complicated environment the advantage of remote sensing monitoring,highlight the characteristics of SAR image is constructed the texture of time series,analysis the relationship between the response of crops growth process.The methods of crop planting type discrimination and phenological information extraction were developed to realize the monitoring of farmland planting status and extraction of planting system,which provided theoretical and method support for the innovation and development of precision agriculture information technology in the complex environment of karst mountainous areas. |
PDF Full Text Request