Study On Cotton Yield Online Monitoring System Based On Photoelectric Sensing

Precision agriculture technology has become an important researchissue for many countries in this new century. The crop yield is anintegrated reflection of the influence to crops given by farmland spatialand temporal difference. Achieving production space distributioninformation is the basis of farmland variable operation. It has an importantsignificance to get quick and accurate access to yield distributioninformation for judging the economic benefits, carrying out the precisionfarming, improving the efficiency of agricultural operations. So far cropyield monitoring technology has been researched and carried out in foodsuch as corn and wheat in domestic, however still in its infancy for cottonharvest. Cotton is an important cash crops and strategic materials and itspicking is a seasonal production process of large and high intensity labor.With mainly picked by hands, China’s cotton meet the problem of lowlevel mechanization throughout a large area and high labor costs, whichhas seriously hampered its competitiveness on the international market andthe steady development in domestic.In response to these issues, a cotton yield online monitoring systembased on optical frequency conversion principle was developed under theimplementation of the national863agricultural special subject of researchon cotton picker intelligent control system. The system designed by theideas of bus mainly includes three parts: host computer system for thegraphical man-machine interface and integrated information processing,flow rate sensor system for real-time measurement and other auxiliary sensor system.Study was carried out on the working of flow rate sensor based onphotoelectric principle and embedded control systems via experiments.Results show that the optical method has the advantages of high resolutionand robustness. An industrial embedded SBC has been used as the corehardware of the PC control system. Virtual instrumentation and graphicalprogramming language was applied in developing the control system andhuman-machine interface, making the calculation of yield model,production information display and system settings. Designing system withCAN-bus and communication protocol provide reliability and standards forsystem integration and node expansion. The introduction of GPS/GPRSmodule is mainly used for acquisition of farmland DGPS positioningsignal and wireless communication with the remote server. Besides, a yieldprediction model of neural network based on genetic algorithms wasbuilded to meet the non-linear and multi-input characteristics of the system.Calibration experiments were designed to train and validate the model. Thegood accurate prediction results show the ability of meeting systemdemands.