民用航空器注册号识别的鲁棒性方法研究

为了提高空管自动化水平并降低运行成本,提出一种基于计算机图像处理技术的民用航空器注册号识别方法.首先将全局阈值分割算法与形态学分割算法结合,实现对航空器图像中注册号的精确定位.然后采用主成分分析法解决注册号倾斜校正问题,并分别针对粘连和非粘连情况的注册号字符进行有效切割.最后利用基于特征提取的模板匹配算法,完成对注册号字符的准确识别.在此基础上,基于MATLAB搭建实验测试平台对上述方法进行仿真验证.实验结果表明,方法能够针对不同角度拍摄的航空器图像快速、准确地识别其注册号,单幅图像识别耗时在0.5 s左右.相比于其他方法,在图像拍摄方面要求较低,有效性与鲁棒性较高,计算速度较快,可以满足实时监控的需求.

A Robust Method for Civil Aircraft Registration Number Recognition

In order to improve the automation level of air traffic management and reduce operating cost,a registration number recognition method for civil aircraft based on computer image processing technique is proposed.Firstly,the method which combines global threshold segmentation algorithm and morphological segmentation algorithm,realizes the precise localization of registration number from aircraft image.Secondly,the skewness correction of registration number is solved by Principal Component Analysis method.Then the effective cutting of registration number characters is implemented both for adhesive and separated situations.Finally,the pattern matching algorithm based on feature extraction is applied to identify the registration number characters correctly.Moreover,the experimental test platform based on MATLAB is established in order to verify the proposed method.The experimental results demonstrate that the proposed method recognizes the registration number rapidly and correctly for different images.It costs about 0.5 seconds for single image recognition.Compared with other methods,the proposed method doesn′t need high-level capture image.The high efficiency,robustness and computation speed meet the requirement of real-time monitoring.