【摘要】 针对现阶段中华绒螯蟹(Eriocheir sinensis)防伪技术准确性差和可靠性低的问题,本文提出一种基于图像匹配的蟹背壳特征识别算法。首先,分割出具有唯一性的中华绒螯蟹背壳图像,用SURF算法提取特征点,然后使用双向FLANN算法与数据库中保存的特征点匹配,以匹配点与最近欧式距离和次近欧式距离匹配点的位置关系作为误匹配检测的依据,采用正确匹配对的占比验证其相似度,以此实现河蟹单体的识别。研究表明,在近似的拍摄条件下,两幅属于同一只蟹的背壳图像相似度可达到0.7以上,而属于不同蟹的背壳图像的相似度则偏低。本研究方法可有效解决目前使用绑在螯足上的二维码或条形码等标识物作为中华绒螯蟹商品蟹溯源及其防伪技术准确性和可靠性差的问题,为商品蟹的单体识别与精准溯源提供一项技术手段。更多还原

【Abstract】 Aiming at the poor accuracy and lower reliability of current anti-counterfeiting methods of Chinese mitten crab Eriocheir sinensis, This paper suggests a carapace feature recognition algorithm for Chinese mitten crab based on image matching. Firstly, the unique carapace image of Chinese mitten crab is segmented on which feature points are extracted by means of SURF(Speeded up robust features), then these feature points are matched up with those stored in the database by virtue of the two-way FLANN(Fast library for approximate nearest neighbors) algorithm. According to the positional relationship between these matching points, the nearest Euclidean distance’s and the second nearest Eucli-dean distance’s matching points, it is used as the basis of mismatch detection. And the ratio of the correct matching point pair is used to verify its similarity so as to realize the individual identification of a crab. This findings show that the similarity of the two carapace images belonging to the same crab is higher than the threshold value of 0.7 and lower for other two different crabs under the similar conditions of image acquisition. This research can solve the problem of poor technical accuracy and low reliability of a crab’s traceability and anti-counterfeiting using barcodes or two-dimensional barcodes attaching to a crab’s chelae, and can contribute a technique for dealing with accurate individual identification of merchandise crabs to realize the purpose of anti-counterfeiting and Accurate traceabi lity.更多还原