Machine vision-based citrus yield mapping system

The variability of yield in citrus groves is important for growers to know to make correct management decisions. Current citrus yield mapping systems require hand harvesting which is labor intensive. In computer vision-based agricultural applications for yield mapping, detecting occluded and non-occluded fruit from acquired images of trees is one of the major problems. Since there are no completely robust and efficient methods, detecting occluded fruit from acquired images has received much attention in computer vision-based agricultural applications. This paper presents an automatic machine vision system with two charge coupled device (CCD) cameras, ultrasonic sensors, an encoder and a differential Global Positioning System (GPS) receiver to estimate citrus yield. An alternative computer vision algorithm was proposed to recognize visible and partially occluded citrus fruit from trees. The average fruit size was determined from images using ultrasonic sensors measuring a distance between the cameras and the fruit laden trees. Finally, a citrus yield map was created to show yield variability for site-specific crop management.