摘要
作为实现果蔬采摘作业自动化的关键难题之一,田间环境下的果蔬快速识别与定位受到了广泛的关注。综述了常见的果蔬识别硬件系统和识别算法;总结和分析了果蔬图像预处理、颜色特征选择、图像分割算法设计及图像后处理的研究;综述了常见果蔬识别算法、成簇果蔬识别算法及被遮挡果蔬识别算法的研究;分析和比较了果蔬定位中常用的有源测距法和无源测距法,以及果蔬无源测距中常见的立体匹配算法。分析了田间环境下果蔬识别和定位研究存在的主要问题,并展望了发展趋势。
As one of the key problem for the robotic harvesting, the real-time recognition and localization of fruits and vegetables in field have drawn wide attention. First, the development of common hardware and algorithms for the recognition of fruits and vegetables was reviewed. For the image segmentation of fruits and vegetables, researches on image preprocessing, color feature selecting, image segmentation designing and image post processing were summarized and analyzed. For the recognition of fruits and vegetables, researches on common algorithms, recognition algorithms for clustered objects and for occluded objects were reviewed. Second, active and passive range finder methods which have been used commonly in the localization of fruits and vegetables were analyzed and compared. Moreover, stereo matching methods used in the passive range finder methods for fruits and vegetables were analyzed and compared, too. Finally, existing problems in researches on the recognition and localization for fruits and vegetables in field were analyzed, and research prospects were also presented.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2013年第11期208-223,共16页
Transactions of the Chinese Society for Agricultural Machinery
基金
国家自然科学基金资助项目(61105024)
'十二五'国家科技支撑计划资助项目(2011BAD20B12)
关键词
田间
果蔬
采摘机器人
识别
定位
Field Fruit and vegetable Harvesting robot Recognition Localization
