应用自回归模型的条码图像盲去模糊算法

Barcode Image Blind Deblur Algorithm Based on Auto-Regressive Model

条码的解码景深是条码解码能力的重要评估指标。解码景深越长,对模糊条码的识别能力要求越高。该问题是一个典型的盲去模糊问题。本文结合条码的二值化特性,基于自回归模型,设计相应的优化函数和优化方法,提出条码图像盲去模糊算法Blind-TVDWSAR。通过仿真实验和实拍图像实验验证了所提算法的有效性。将该算法嵌入条码识读设备发现,相比未做盲去模糊,条码解码景深有至少5cm的延拓效果。

The field depth of barcode-decoding is an important evaluation index for barcode recognition, which is a typical blind deconvolution problem. The further the decoding depth of field, the greater the requirement of barcode recognition. Based on total-variation and double well regularity and simultaneous auto-regressive model, this paper proposes a barcode image blind deblur algorithm named Blind-TVDWSAR. The effectiveness of the proposed method is verified by simulation and actual experiments. By embedding this algorithm into the barcode-recognition device, it is found that the field depth of barcode-decoding has a continuation effect of at least 5 cm.