Key technologies in apple harvesting robot for standardized orchards: A comprehensive review of innovations, challenges, and future directions

Abstract

• A comprehensive research assessment was conducted on fresh market apple harvest robotics. • Deep learning performs better than traditional handcrafted feature algorithms in apple detection. • Multi-source sensor fusion provides apple localization information accurately. • Multiple arm and vacuum end-effectors require more attention in future research. Harvesting robots offer a promising solution for the automated collection of fresh apples, significantly enhancing efficiency and productivity in the agricultural sector. However, their effectiveness in orchards is influenced by various factors, including orchard production models, environmental conditions, and sensor technologies. This study aims to enhance the intelligence and commercial viability of harvesting robots by summarizing, analyzing, and forecasting critical technologies that influence robot performance in areas such as fruit identification, localization, arm motion planning, end-effectors design, and fruit detachment mechanisms. A comprehensive review of apple recognition methods is presented, which introduces various sensor localization technologies, examines robotic arms and motion planning algorithms for harvesting, and analyzes end-effectors and their corresponding detachment methods. An overview of the current state of apple-harvesting robots is provided, highlighting those that are poised for commercialization. Significant challenges in the field are addressed, including lengthy harvesting cycles, poor success rates, severe fruit damage, and high operational costs. The paper also outlines promising future directions for advancing commercial applications in this domain. These include innovations aimed at simplifying orchard environmental complexities, optimizing recognition and motion planning algorithms, integrating diverse sensor technologies, enabling autonomous multi-arm collaboration, developing vacuum-based end-effectors, and implementing secondary picking strategies. Collectively, these advancements are geared towards the creation of market-ready apple harvesting robots, ultimately fostering growth within the apple industry.