Researches And Applications Of The Instrument Images' Input And Pretreatment

We can get the reading information of the instruments by using machine vision to recognize the readings of the instruments' images and then pick-up and recognize the target place without change the primary arrangements, which is called the auto recognition of the instrument images and it is the hotspot of the research now. But there are few reports about how to assess the performance of recognize arithmetic fast and expediently. If the computer knows the exact readings of instrument images in advance, it can sort the rates of the images which had not been recognized or were recognized wrong automatically. There are two methods to let the computer know the readings of the instrument images in advance: one is building up an instrument display system whose display values are controlled by the computer program and then uses the USB camera capture the images of the instruments; the other is using the computer program to emulate the instrument images. For the first method, we should do some pretreatments to wipe off the noise of the images. This paper aims at such a goal, and made the following contributions:1. Achieved methods to input the images to the image processing system. In order to sort the rates of the images which had not been recognized or were recognized wrongly.(1). Capturing the images by the USB camera in high efficiencyIn order to integrate the system easily and use the resource of the system fully, it developed a capture program in the VC++ environment used the VFW method in this paper. Added the multithread in the capture program, in such ways we can use the CPU fully and advance the full performance of the entire system.(2). Instrument display controlled by computer programThe computer controlled the instrument display module by the serial port. It includes the display value and the time between the changes of the two consecutive values. The display numbers were saved in the computer in order to contrast with the recognize results in the future, so as to sort the rates of the images wich had not been recognized or were recognized wrongly.(3). Instruments' images emulated by the computer:Building up the geometry emulate model of the instruments' display, using the GDI to draw the instrument display emulating images, wiping off the glittering by using the double buffering, emulating a PID controller to control the display value of the emulated instruments, and changing the rand numbers which were short to the float style.2. The pretreatment of the instrument images:We should wipe off the background noise before the automatic recognition. It used the color model methods to wipe off the noise of the nixie light display instruments' noise, in such way we can improve the speed of the image processing and the commonality of the arithmetic. And it enhanced the contrast of the LCD display instruments' image by wavelet liftering enhancement, and its effect was ideal.3. Programmed all the arithmetics brought by this paper and the results had utility values.All the instrument images input and pretreatment methods had been used in the image processing system. Programmed the corresponding subprograms. The functions' effect were very well and it will be brought forth in the corresponding chapters.