摘要
双目视觉系统广泛应用于现代物流,其中高效精准的匹配定位算法是实现快速有序物流作业的关键技术基础。为缩短特征点匹配时间,提高匹配准确率,提出了一种基于加速稳健特征的自适应阈值匹配定位算法。首先采用加速稳健特征算法分别对左右图像提取特征向量,应用自适应双阈值最近邻法获取初始匹配对;然后在极线约束条件下筛除误匹配对;最后采用距离特征判别法与角度特征判别法进一步剔除误匹配对。条烟定位实验显示该算法的平均匹配正确率可达90%以上,且定位时间较加速稳健特征算法最多可减少40%,实验结果表明了所提算法具有高效准确的优势,对自动码垛系统具有一定的实际应用价值。
Binocular vision system is widely used in modern logistics, where efficient and accurate matching positioning algorithm is the key technological foundation to achieve fast and orderly logistics operations. In order to shorten the time of feature point matching and improve the accuracy of matching, this paper proposes a speeded up robust feature based adaptive threshold matching positioning algorithm. Firstly, the SURF algorithm is used to extract the feature vectors of the left and right images, and the adaptive double threshold nearest neighbor method is used to obtain initial matching pairs. Then, under the condition of polar line constraint, the mismatching pairs are screened out. Finally, the distance feature discrimination and angle feature discrimination methods are adopted to further eliminate mismatching pairs. The cartoon cigarette positioning experiments show that the average matching correct rate of the proposed algorithm can reach 90% above, and the positioning time of the algorithm can be reduced by 40% at most, compared with that of the speeded up robust feature algorithm. The experiment results demonstrate that the proposed algorithm has the superiority of high efficiency and accuracy, and thus has certain practical application value to automatic palletizing system.
作者
罗久飞
邱广
张毅
冯松
韩冷
Luo Jiufei;Qiu Guang;Zhang Yi;Feng Song;Han Leng(Robotics and Advanced Manufacturing Research Center,School of Advanced Manufacturing Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2020年第3期240-247,共8页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(51705059)项目资助
关键词
双目视觉
极线约束
图像匹配
加速稳健特征
binocular vision
polar line constraint
image matching
speeded up robust feature(SURF)
