Electrical Safety Testing Embedded Systems, Digital Image Recognition

Based on a set of multi-function intelligent security appliance, this text redesigned the hardware and software on the part of Low voltage startup and Power testing ,not only condensed old hardware circuit but enhanced the stabilization of the hardware circuit, and made the testing precision of voltage, current and power no more than 2%o measure, designed interlock protect of the whole system, and finished the experimental appliance. The main processor was MEG128 made by ATMEL and sub-processor was MEG8 of ATMEL. Due to the sole-designed of the Low voltage startup and Power testing, for combined with the other four modules, we built the hardware and software flats, it was also convenient for continued extension. But, if there was different system wanted to combined in this system, we must redesigned all the hardware and software, it was a very hard work and could bring incompatibility. For this instance, the text designed a suit of number discriminating embedding system. The system faced to different appliance. The combination was not electric connection, namely, it got testing signal from display screen by CCD camera and Ezcapture image capturing card, then carrying the signal to old system. So it was a new system and keeping old system untouched. It also was convenient for the appliances that had not standard interface or was incompatibility to buildup a new system fast.The text used CS5460 chip as main chip of measurement scheme, found a new date sampling way. When designed the hardware, we improved measure precision at the same time reduced hardware circuit. When designed software, we used MEG8 as process chip, measure the power, voltage and current fast and exactly. So we advanced the precision and real time characteristic of measurement.The pivotal part of this text was the process and identification of numeral image. The image we got was not standard caused by beam condition and yawn and so on, we must process them before identified. The process of image treatment is as follow: filter, enhance, making binary image by adaptive threshold and so on. We used the arithmetic of character-suit to get exact image signal. At last we had discussed revised identification about lack-strokes numeral image.