Precise recognition for automated harvesting and grading of strawberries

草莓自动化采收、分级精准识别

• 批准号: LP150100658
• 负责人: Prof Yongsheng Gao
• 金额: $ 28.25万
• 依托单位: Griffith University
• 依托单位国家: 澳大利亚
• 项目类别: Linkage Projects
• 财政年份: 2015
• 资助国家: 澳大利亚
• 起止时间: 2015-12-03 至 2018-12-02
• 项目状态: 已结题
• 来源: https://dataportal.arc.gov.au/RGS/Web/Grants/LP150100658
• 关键词: Precise; recognition; automated; harvesting; grading

This project aims to improve automated strawberry harvesting to enable industrial harvesters to be deployed for commercial use and to lift the productivity of the Australian fruit industry. Precise recognition and grading of strawberries is a major obstacle in developing fully-automated commercial strawberry harvesting systems. Current colour-based fruit recognition techniques have intrinsic limitations in meeting the needs of automatic strawberry harvesting. This project aims to investigate high-level syntactic recognition approaches that embed high-order texture patterns of ripe fruit and hyperspectral analysis techniques to achieve partially occluded fruit recognition and grading of fruit at the level required by commercial production.

该项目旨在改进自动化草莓收获，使工业收获机能够用于商业用途，并提高澳大利亚水果产业的生产力。草莓的精确识别和分级是开发全自动商业草莓收获系统的主要障碍。目前基于颜色的水果识别技术在满足草莓自动收获的需要方面存在固有的局限性。本项目旨在研究嵌入成熟水果高阶纹理模式和高光谱分析技术的高级句法识别方法，以实现部分遮挡水果的识别和商业生产所需的水果分级。