Machine Vision for Automated Corn Plant Spacing, Growth Stage and Population Measurements – Part II: Plant Identification

After the real-time image-sequencing process, a set of individual corn plant and plant stem center identification algorithms were developed and implemented with a highly integrated software environment. An average corn plant spacing measurement error of less than 10 mm was achieved with minimal manual corrections. In addition, for accurate identification of corn plants, weeds must be differentiated from crop. Algorithms for this purpose, such as the robust crop row detection algorithm using M-estimates, have potential in other precision agricultural operations, e.g. selective weed control and guided cultivation.