【传统CV】Anovel adaptive morphological approach for degraded character image segmentation

最新推荐文章于 2024-10-16 19:08:56 发布

艾尔_1222

最新推荐文章于 2024-10-16 19:08:56 发布

阅读量325

点赞数

分类专栏：字符分割文章标签：算法计算机视觉深度学习

本文链接：https://blog.csdn.net/m0_37661841/article/details/103582758

版权

本文提出了一种基于数学形态学（MM）的自适应图像分割和提取方法，用于处理退化的字符图像，特别是车牌信息。通过结合启发式方法确定潜在的分割点，算法能解决重叠、连接和碎片化字符的问题。论文详细介绍了从二值化、去噪到字符自适应分割的步骤，并应用于实际的车牌图像，以识别和定位每个字符。此外，还探讨了如何利用几何信息和人工神经网络进行模式识别。

摘要由CSDN通过智能技术生成

【传统CV】Anovel adaptive morphological approach for degraded character image segmentation

基于mathematical morphology的图像分割和提取；作者的论文主要用来提取车牌信息；

Introductin

In this paper, a new adaptive segmentation and extraction approach is proposed based on mathematical morphology(MM) operators in conjunction with heuristics that determine the potential segmentation points. The license plate digitized images, that represent the degraded images, are acquired from real scenes of moving vehicles taken by cameras installed at the streets [7] in Uberlândia city. Fig. 2 shows a sample of these digitized images. The images are often acquired under dubious light and focusing conditions, and are often badly degraded images that include serious problems such as fragmented, overlapped or connected characters in their corresponding binary images. Examp