Autonomous navigation for a wolfberry picking robot using visual cues and fuzzy control
Yue Ma,Wenqiang Zhang,Waqar S. Qureshi,Chao Gao,Chunlong Zhang,Wei Li
DOI: https://doi.org/10.1016/j.inpa.2020.04.005
2021-03-01
Information Processing in Agriculture
Abstract:Lycium barbarum commonly known as wolfberry or Goji is considered an important ingredient in Japanese, Korean, Vietnamese, and Chinese food and medicine. It is cultivated extensively in these countries and is usually harvested manually, which is labor-intensive and tedious task. To improve the efficiency of harvesting and reduce manual labor, automatic harvesting technology has been investigated by many researchers in recent years. In this paper, an autonomous navigation algorithm using visual cues and fuzzy control is proposed for Wolfberry orchards. At first, we propose a new weightage (2.4B-0.9G-R) to convert a color image into a grayscale image for better identification of the trunk of Lycium barbarum, the minimum positive circumscribed rectangle is used to describe the contours. Then, using the contour points the least square method is used to fit the navigation line and a region of interest (ROI) is computed that improves the real-time accuracy of the system. Finally, a set of fuzzy controllers, for pneumatic steering system, is designed to achieve real-time autonomous navigation in wolfberry orchard. Static image experiments show that the average accuracy rate of the algorithm is above 90%, and the average time consumption is approximately 162ms, with good robustness and real-time performance. The experimental results show that when the speed is 1km/h, the maximum lateral deviation is less than 6.2cm and the average lateral deviation is 2.9cm, which meets the requirements of automatic picking of wolfberry picking robot in real-world environments.