Application method of performing reading identification for communication cabinet through Raspberry Pi

• Main Authors: CHEN YILONG, LI WEIHUA, CHEN XIAOLU, GONG WEI, XIAO YUNJIE, XIA SHIJUN
• Format: Patent
• Language: Chinese; English
• Published: 31.07.2018
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses an application method of performing reading identification for a communication cabinet through the Raspberry Pi. The application method comprises the following process that step S1, the Raspberry Pi acquires the cabinet image of each communication cabinet; step S2, the cabinet reading area is identified and separated from the cabinet image; step S3, reading result contour detection is performed on the cabinet reading area obtained by separation, the reading number area of the reading area is split into single number images, and the obtained single number images are identified through the network; and step S4, the mobile terminal performs the tensorflow program step and identifies the split single number images by using the pre-trained mnist model. The application method has the advantages of enhancing the reading identification speed and accuracy of the communication cabinet. 本发明公开了种利用树莓派对通信机柜进行示数识别的应用方法，包含以下过程：步骤S1、树莓派获取每个通信机柜的机柜图像；步骤S2、分别从机柜图像中识别并分离出机柜示数区域；步骤S3、分别对分离出的机柜示数区域进行读数结果轮廓检