【摘要】 提出了一种基于模板匹配和神经网络的X射线底片编号、定位分割和识别方法,可以有效的检测和识别底片中的低质量和模糊的编号字符,识别率可达92%,识别时间不超过1s.更多还原

【Abstract】 A serial number location segmentation and recognition method based on template matching and neural network in X-ray negative is proposed. The low-quality and blur serial number character in X-ray negative can be inspected and recognized effectively and the recognition ratio can come to 92 percent and the process is less than 1 s.更多还原