Real-time swath width sensing for grain combines

The practice of precision agriculture is utilized in grain production to make decisions to lower costs and increase profits. To make these decisions, an accurate determination of yield is required at each position within the field. Current yield measurement incorporates yield monitors and Global Positioning System (GPS) receivers to calculate yield from grain flow and position. The primary problem encountered is a varying width of cut due to operator error or when cutting "point rows" near ends and edges of a field. To more accurately measure yield, a method for measuring harvest cutting width has been developed. The non-intrusive sensing system consists of infrared emitter-detector pairs mounted along the length of the combine header reel. Each sensor element emits a beam of light into the crop canopy and detects the signal as it is reflected by plants. The emitted signal modulated pulses of light that are detected when a crop is present. Zero pulses from a detector indicate no plants were present; a higher number indicates plants were in the field of view. The sensors are active once each reel revolution. Software analysis was developed and performed to assemble the recorded numbers of pulses counted into a data matrix and to use this information to predict other locations that should have detected plants. The software then determined the measured swath width. The sensing system was tested on a combine harvesting with a 12 ft header. Field testing was then performed in production soybeans. Actual detection by the sensors occurred at only 16% of the locations in the harvest area where plants were present. However, the software analysis routines increased the locations plants were present to 62%. Data analysis routines based on knowledge of the crop canopy increased accuracy significantly. Refinements are necessary in the sensor elements to increase their sensing range and to improve the location at which sensors are sampled. Increasing the detection by the sensors allows the software to more accurately determine the harvested width. Future testing should focus on maximizing the efficiency and detection rate of the sensors under varying crop conditions.