针对传统的物流编号都是以二维码为基础的识别,这种方法一旦遇到号码脱落等情况会影响货物识别的准确性.为了解决这一问题,提出一种基于少量局部HOG特征的稀物流编码识别算法.方法选取了多个物流编码中的归一化像素特征将其转换为特征向量,构建识别关联特征库,对特征表达下的少量关联特征进行系数计算,完成编码识别.实验结果表明,与传统的电子标签识别方式相比,该方法不仅有很高的正确识别率,而且对于有部分缺失的物流编号也能完成有效识别,有很大的应用价值.%In this paper, we proposed a logistics code identification algorithm based on small-quantity local HOG characteristic. More specifically, we selected the normalized pixel characteristic of a number of logistics codes, transformed it into feature vectors and established the database of correlated identifiable features. Then we calculated the coefficient on the basis of the correlated small-quantity features and finished the identification of the codes. An experiment showed that, compared with traditional e—label identification methods, this method not only had higher correct identification rate, but also could effectively identify logistics codes that were partial missing.
展开▼
