摘要
在农业果实自动化采摘过程中,果实检测易受光照、尺度变换等影响导致算法缺乏泛化性.针对这些问题,本文提出了一种基于条件增强随机蕨(conditional boosted random ferns,CBRFs)的果园苹果检测方法.首先,基于面向HOG(histogram of oriented gradients)特征域的增强随机蕨分类器(boosted random ferns,BRFs),在选取特征时限制特征方向区间选择的随机性;其次,改变分类器的二进制特征产生方式,以避免从图像平坦区域中随机选取的特征出现误分类的情况;最后,在自制苹果检测数据集上进行分类器训练,构建条件增强随机蕨分类器,使其适应于果实重叠、枝叶遮挡和光照不均等复杂环境下的果园苹果检测.在测试集上的验证结果表明,F1值与原增强随机蕨模型相比提高了5.44%,这说明本文所提方法能有效增强果园中苹果的检测效果.
In the process of automatic picking of agricultural fruits,fruit detection is easily affected by light and scale transformation,resulting in the lack of generalization of the algorithm.To solve these problems,this paper proposes an orchard apple detection method named conditional boosted random ferns(CBRFs).Firstly,on the basis of the boosted random ferns(BRFs)for histogram of oriented gradients(HOG)feature domain,the randomness of feature direction interval selection is limited while selecting features.Secondly,the binary feature generation mode of the classifier is changed to avoid misclassification of the features randomly selected from the flat area of the image.Finally,the classifier is trained on the self-made apple detection data set,and the conditional boosted random ferns classifier is constructed to make it more suitable for orchard apple detection in complex environment of fruit overlap,leaf occlusion and uneven illumination.The results tested on the test set show that the F1 value is increased by 5.44%compared with the original boosted random ferns model,which shows that the proposed method can effectively enhance the detection effect of apples in the orchard.
作者
白金华
张乾
范玉
张宇航
何兴
BAI Jinhua;ZHANG Qian;FAN Yu;ZHANG Yuhang;HE Xing(School of Data Science and Information Engineering,Guizhou Minzu University,Guiyang,Guizhou 550025,China;Academic Affairs Office,Guizhou Minzu University,Guiyang,Guizhou 550025,China;Key Laboratory of Pattern Recognition and Intelligent Systems of Guizhou Province,Guiyang,Guizhou 550025,China)
出处
《湖南城市学院学报（自然科学版）》
CAS
2022年第1期62-66,共5页
Journal of Hunan City University:Natural Science
基金
贵州省教育厅自然科学研究项目(黔教合KY字[2018]141)。
