Vision-based navigation and guidance for agricultural autonomous vehicles and robots: A review

• Published in: Computers and electronics in agriculture 2023-02, Vol.205, p.107584, Article 107584
• Main Authors: Bai, Yuhao, Zhang, Baohua, Xu, Naimin, Zhou, Jun, Shi, Jiayou, Diao, Zhihua
• Format: Article
• Language: English
• Subjects: Agricultural robots; Autonomous vehicles; Computer vision; Navigation; Path planning; Precision farming; Smart agriculture
• ISSN: 0168-1699; 1872-7107
• DOI: 10.1016/j.compag.2022.107584

•A detailed summary about the research status of vision-based navigation technology for agricultural robots was given.•Vision sensors characteristics and vision information processing methods were summarized.•The applications of vision-based navigation methods in robotic agricultural tasks were summarized.•Challenges and future trends of vision-based navigation technology in agricultural robots are reported. Autonomous navigation of agricultural robots and vehicles in agricultural environments is a prerequisite for the accomplishment of various tasks. However, precision navigation of agricultural robots is still a challenging issue due to the complex and unstructured nature of the agricultural environment. With the development of electronics and information technology, machine vision technology has become a promising tool for real-time and accurate navigation of agricultural robots. Due to low hardware cost and rich visual information, machine vision technology has been intensively studied and widely used in the field of autonomous navigation agricultural robots. Aiming at the special complexity of agricultural environment, this paper reviews the research advances of agricultural autonomous vehicle and robot navigation and guidance based on machine vision. A detailed summary of the development and characteristics of various vision sensors and systems is given. Key visual navigation information processing technologies such as filtering-based data computation, segmentation-based data computation and line-detection-based data computation for agricultural robots are discussed. Special attention is paid to the application of vision-based navigation technology for agricultural robots, i.e., environment perception and mapping, robot localization, and path planning. Finally, the challenges of machine vision in agricultural robot navigation are discussed, and the future development of vision sensor technology and autonomous navigation technology is prospected.