Design And Implementation Of Solar Array Image Analysis Software

Solar array is the only source of Satellite energy, power installations issolar cells which was broken Easily, the damage of solar cells are vital foroutput power of solar array. There are strict requirements of it's solar cells'sdamage rate such as crack or scratch on the relevant norms of solar cells inspace. In the process of developing the satellite, after transport, mechanicalexperiments and thermal cycle test of the solar array , it may lead to thedamaged solar cells , therefore, the need for a Solar Array Detection to ensurethat its damage rate was in line with the technical requirements. The currentdetection method is eye method, measurement using the solar microscope for avisual inspection, this kind of method is not too exactly, the development of aprecise, rapid detection of solar array panels cracks, defects such as scratchesthe status of automated image analysis software is essential.In this paper, the main problem is how to solar array image of their owncharacteristics and digital image processing computer with the knowledge tocomplete removal of the solar arrays image gate line, noise and determinecracks, scratches, and other technologies. Image processing is to use a series ofspecific actions to change the image pixels in order to achieve specific goals,such as making more explicit images, or images extracted from certainspecified information. Solar Array images are basically BMP image bitmapformat, but the solar array images have their own characteristics andcomplexity of the conventional image processing method is not entirelyapplicable, So based on the solar arrays image of their own characteristics, thepaper and software proposed and adopted a number of special methods forhandling binary and noise to specific image manipulation.(1)Solar Array image characteristics and particularities:Solar array has alot of information and images interference , Generally includes the gridlines (horizontal and vertical), and the mist is dust stains or information, size, shading from the noise, these may be big noise dust, soil particlesor droplets traces;(2)Image pixel "peak" information: The so-called image pixel "peak"information refers to the image pixels in the horizontal or vertical pixelvalue from a low starting point to the highest point value of the pixelsand then to the lowest point in a chronological order, the peak pixelprocessing and identifying the solar arrays used by the imageprocessing one of the major means. Such laws passed pixels we candetermine the composition of pixels is scattered or non-linear ornon-availability of information;(3)Decontamination (dust) treatment: Image stains on the mist or dust canalso be counted as a kind of noise, the more noise is such a meticulousand cluttered, the first such general dust mist treatment to "dilution" orremoval. General approach is in the horizontal or vertical images onthe scan, the peak value of a pixel is not obvious from veryconcentrated and should be removed region;(4)Binarization: Binary Image processing is the image of the points on 0or gray home to 255, that is, the whole image of a clearblack-and-white effect. This paper proposes a "Star peak" ofbinarization: The basic idea of this method is to "peak", but this is notthe peak horizontal or vertical, but to judge a pixel of the fourdirections, but is not to judge the direction of all four of the points, justjudgement a pixel around four points (below) can be, so all of a pixelimages to be scanned, a point on each judge around four points ofpixels, if the conditions are consistent with judge that is or is not a"peak" point. Specific in the judgement, the pixel values greater than128 points, can be considered a bright spot, if its not less than fourpoints around the value of its own and will be around four points at thesame time home to 255, just like it around The pixel "light", less than 128 points for the "light" two or three points, this approach is like thestars lit the sky around the same, so called "Star peak" Binary ;(5)Erase the gate lines: Gate line is the image of the strip-line, in thecalculation of specific, hough transform View images used in thehorizontal or vertical line to erase;(6)Erase noise: This paper summarized the tests to more than four typesof noise, and in accordance with the noise characteristics of a "arearatio" method of noise erase: Because the images are scratches orcracks form of a straight line, if not completely linear form but also alot of straight line or curve, a straight line or curve of the image can beclear if the identification, then these images itself and its proportion tothe number of pixels the largest area ratio are small (between 0.3-0.5),the largest so-called occupied area is the level of their maximum lengthand vertical maximum length of the product. Such a judgement methodcan remove most similar to a square or circular area of the noise andsome irregular noise. Finally, the proposed amendment to the noisemethod.Generally speaking, a solar array image to go through decontamination(dust) with the value of grid lines and erase borders and to deal with suchnoise, and end up with cracks or scratches of the two plans.Design and Implementation of the system, the software system useshierarchical structure design: First layer: basic image processing functions ofthe second layer: Object-based image; third layer: data processing layer;Fourth layer: functional layer; fifth: interface. Low-level functions to providehigh-level support, concrete realization of senior officials. SoftwareEngineering is a major ideological analysis, design and realization of solararray image analysis software for the preparation, adoption of digital imageprocessing technology and related image processing techniques and algorithmsto achieve removal grid lines, to the noise, judge cracks, scratches information storage battery-damaged location, print and display the results, and otherfunctions, procedures to the needs of image processing; Well-designed userinterface, easier to use, and a higher degree of automation and establish asound file system (Figure UML), object-oriented software more in line withthe norms, and further to achieve its reusability and extensibility.This paper introduced the final of the software interface and the actualprocessing and gives examples of the conclusions and inadequate, and inrelation to content in the future work of the need to further explore the issue.