MODELING TRENDS IN CROP YIELDS

Abstract A computer implemented method: sending requests to remote sensors installed on agricultural equipment to provide data comprising yields of crops harvested from agricultural fields at multiple time points; receiving and analyzing the data, receiving a request to generate particular yield data: using yield dependency, determining factors that impact the yields where factors are time independent and represents spatial dependencies between the agricultural fields; decomposing the yield data that identifies dependencies according to the said factors; using data approximation to generate, based on the decomposed yield data, the particular yield data using a modified singular value decomposition that uses at least said factor; wherein the particular yield data includes fewer values than the yield data; using data reconstruction to generate forecasted yield data using the particular yield data into the yield data; wherein the generating forecasted yield data by processing and incorporating the particular yield data, which includes fewer values than the yield data, uses less computational and storage resources than determining forecasted yield using the yield data; transmitting results to enable operational decisions in controlling an agricultural vehicle for crop planting, fertilizing or harvesting. WO 2017/136417 PCT/US2017/016007 Fig.1 .. Feielo Data External Data 1_15 Cab 111 Agricultural 113 Computer Apparatus i+ 1_12 Remote Seno 114 Application Controller 1Q9 Network(s) 132 Communication Layer 180 Code Instructions 1182 Data Receiving Instructions 83 Data Analyzing Instructions 184 Yield Dependency Instructionsa 16 Model Data 18 Data Decomposition Instrucin Field Data Repository 186 Data Approximation Instructions 17Data Reconstruction Instructions 188 Results Interpretation Instrucin - W 134 Presentation Layer 140 Data ManagementLae 150 Hardware/Virtualization Layer 130 Agricultural Intelligence Computer System