Computer vision to locate fruit on a tree

Abstract

ABSTRACT RESEARCH on the application of computer vision as a sensing system for robotic harvesting of fruit was performed. A vision system was developed to collect images. Testing was conducted during the day and at night on peach and apple trees. A computer algorithm was developed to enhance and extract information from images. This extracted information was used to classify objects within these images as either **multiple fruit, ^'single fruit, or '*noise. The algorithm correctly handled approximately 92% of the objects and identified approximately 89% of the fruit found in images obtained at night. Computer vision was determined to be a viable technique for detecting and identifying fruit on trees..