Skip to main content
ARS Home » Southeast Area » Houma, Louisiana » Sugarcane Research » Research » Publications at this Location » Publication #353173
Title: Robotics for sugarcane cultivation: analysis of billet quality using computer vision

Author
item ALENCASTRE-MIRANDA, MOISES - Massachusetts Institute Of Technology
item DAVIDSON, JOSEPH - Massachusetts Institute Of Technology
item Johnson, Richard
item WAGUESPACK, HERMAN - American Sugar Cane League
item KREBS, HERMANA - Massachusetts Institute Of Technology

Submitted to: International of Electrical and Electronics Engineers (IEEE) Robotics and Automation Letters
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/16/2018
Publication Date: 7/18/2018
Publication URL: https://handle.nal.usda.gov/10113/6961194
Citation: Alencastre-Miranda, M., Davidson, J.R., Johnson, R.M., Waguespack, H., Krebs, H.I. 2018. Robotics for sugarcane cultivation: analysis of billet quality using computer vision. International of Electrical and Electronics Engineers (IEEE) Robotics and Automation Letters. 3(4):3828-3835. https://doi.org/10.1109/LRA.2018.2856999.
DOI: https://doi.org/10.1109/LRA.2018.2856999

Interpretive Summary: In many parts of the world, sugarcane is planted with machines using billets, which are shorter segments of cane harvested and cut by a combine harvester. In other regions the whole stalk is planted either by machine or manually. When using billets, the harvesting process can damage billets, which introduces pathways for disease, and cause an overall reduction of billet quality. Also, if a grower plants billets with a machine, they will use twice as much cane as compared to whole stalk planting. As a first step towards improving sugarcane production with robotics technologies, this paper presents the analysis of sugarcane billet quality using computer vision. A large sample of sugarcane billets was harvested at a research farm in Schriever, Louisiana. A group of crop scientists and growers then categorized the billets into six classes of damage according to physical features visually evident. To better understand the correlation between the type of damage and sugarcane germination, we planted 120 samples from each class in test plots and then recorded plant emergence rates. A dataset of billet pictures was collected with two types of cameras in both outdoor and indoor lighting conditions. The data was analyzed and a classification procedure was developed that resulted in approximately 90% successful classification of sugarcane billet damage. The methods developed in this research will help growers insure that they are planting healthy cane and allow them to maximize yields and increase profitability.

Technical Abstract: In much of the world, sugarcane is planted in a mechanized fashion using billets, which are shorter segments of cane harvested and cut by a combine harvester. The mechanized harvesting process can damage billets, which introduces pathways for disease, and overall reduction of billet quality. Compared to whole stalk planting with manual methods, growers must approximately double the planting density when using billets. As a first step towards improving sugarcane production with robotics technologies, this paper presents the analysis of sugarcane billet quality using computer vision. A large sample of sugarcane billets was harvested at a research farm in Houma, Louisiana. A group of crop scientists and growers then categorized the billets into six classes of damage according to physical features visually evident. To better understand the correlation between the type of damage and sugarcane germination, we planted 120 samples from each class in test plots and then recorded plant emergence rates. A dataset of billet imagery was collected with CCD and stereovision sensors in both outdoor and indoor lighting conditions. Offline image processing resulted in approximately 90% successful classification of sugarcane billet damage.