Small digital recognition in gravel aggregate production scenario

摘要：

　　The production process of aggregate industrial is monitored by the identification of the digital.In the video,the digital character whose size is less than 12 15(The width of the digital image is 12 pixels,and the height is 15 pixels.) is defined as the small digital.The recognition effect is evaluated by the recognition rate and speed.Because of the small digital of pixels in the video,the numerical recognition algorithm based on template matching is not ideal for the small digital recognition.In order to solve this problem,this paper presents a special number of cascaded recognition process.Firstly,the SVM classifier based on HOG features is used to recognize the number of images,and the special digital of the identification results is based on the two level recognition.Compared with the traditional template matching method,this method can improve the respective average recognition rate of the small digital in production and shipment of two scenarios by 27.36% and 5.05%.

关键词： Small Digital Recognition Dissimilarity Characteristic Cascade Identification Structure Characteristics HOG

作者: Ling Li Kai Liu Fei Cheng Long Li

作者单位: School of Computer Science and Technology,Xidian University,Xi’an,China

会议类型: 国际会议

会议名称: 2016 Internatioanl Conference on Engineering and Technology Innovations(ICETI-2016))2016年工程与技术创新国际会议

会议地点: 武汉

会议语种:英文

页码: 200-208

在线出版日期: 2016-03-25（万方平台首次上网日期，不代表论文的发表时间）