Teaching A Computer To Read: Image Analysis Of Electrical Meters

There exists a vast infrastructure of heritage analog and digital meters installed in commercial and industrial applications. These devices typically have no built-in means of automated reading. Modifying the equipment is not a viable option in many applications. With the low cost of USB digital cameras and the availability of LabVIEWTM VISION, a cost-effective method of reading multiple meters of assorted types can be created. Duplicating the process that a human performs while reading a meter display is daunting. However, this process is simplified by using virtual instruments (VIs), which perform essential functions such as edge, pattern and rotation detection. As part of an undergraduate research project, a computer, using LabVIEWTM Vision, together with a USB digital camera is used to read a digital multimeter (DMM) and an analog watt-hour meter. Circular edge detection, pattern searches, and rotation detection are used to locate dials and segments and to determine their values. Horizontal and vertical edge detection and region of interests (ROI) are used to identify and determine the values of a DMM’s display. The ability to read meters with only minimal human interaction increases accuracy and speed. This feature and the ability to create visual data logging have many uses in educational and industrial applications. This paper presents techniques for identifying and reading meter data. The basics of reference images and their use in image analysis are explored in reading legacy DMM and analog watt-hour meters. Introduction and Motivation Various types of electrical meters are required and utilized to take many types of measurements in commercial and industrial applications. There exists a wide range of installed legacy meters which are capable of performing these measurements. The vast majority of these devices lack the ability to be remotely or locally monitored by computers. The ability to be monitored by a computer is becoming increasingly important as the convenience of remotely monitoring equipment outweighs the cost of visiting each piece of equipment. Modern measurement equipment is certainly capable of transmitting its acquired data either through wires of wirelessly. The infrastructure of existing equipment is slowly being replaced with this type of devices. However, the replacement of the installed base of legacy equipment is a costly and daunting task and may not be practical in many instances. The use of cameras in manufacturing automation is well published and very successful. Most systems employ sophisticated and expensive vision equipment for the control of robotic equipment. With the recent popularity of the universal serial bus (USB), a plethora of formidable image capturing devices have emerged. These devices have very high resolution and very low cost. These features make them very attractive for many data acquisition applications as well. Together with sophisticated yet simple to use LabVIEW software, powerful and elegant instrumentation systems can be created. P ge 11196.2 As part of an engineering technology undergraduate research project at Penn State Berks, a webcam was used to capture images of various electrical meter displays for the purpose of extracting the displayed readings. Both digital (LCD) and analog (dials) were used. The LCD of a handheld digital voltmeter was used for the digital display tests. A typical residential Watthour meter was used for the analog display tests. The webcam was controlled by and the images were analyzed by LabVIEW VISION software.