The Application Research Of GPR In Landmine Detection

Land mine is an explosive weapon laid usually just below or above the ground. It is mostly used to make obstacles, prevent enemy from attacking, kill and wound effective and destroy technic equipment.according to data from database of UN demining. Because of years of wars and conflicts, about millions mines in sixty-four countries aren't eliminated. Casualties are nearly 25,000 people every year which are seriously retarding the growth of countries and refugees of war returning homeland. However about one million mines are used for new local wars every year, causing serious disasters and the international community attaches great importance to remove mines left over from the war, have invested a lot of manpower, material resources and financial resources . In short, either in modern wars or sweeping mine after wars, mine detection occupies a very important position. Its role is irreplaceable. Effectively settling the problem of mines has become the focus of attention of the international community, the difficulty lies in the higher demand of mine-detection technology,especially in the post-war demining process.。The traditional demining method uses nonmetal material, which is very unreliable in exactness, alarm-giving, and detection speed, no longer fitting the demand of modern demining. With the development of timespacio electricmagnetic wave theory, computer theory and microelectronic technology, mine detection radar has got improvements both in equipments, data and image processing and the image processing toward the detection target becomes available. ground penetrating radar will be more widely applied in the detection of mine.This paper studies the data characteristics and target identification of ground penetrating radar. According to the testing aim, a suitable choice of the subject is LTD series of ground penetrating radar.Detection performance of grounding penetrating radar in the detection distance and resolution of major, is the echo data processing to analyze and infer the medium structure land rock characteristics of the mines, then echo signal has non-stationary, non-linear attenuation characteristics. Radar can detect the target containing metal ingredients and also can response to the common invariables within the radar lighting scope. Thus, in the high opposition circumstances, any mine differing from surrounding soil material can be detected. Meanwhile, the super broadline radar detecting with high differentiation rate can classify the target information contained in the echo wave signal, which can lower the wrong alarm giving rate.This paper also sums up some regular means of data processing in ground mine dectection radar. Through collected data from field, radar processes the data and obtain the image or signals that can assist in explanation. The aim of data processing is to shadow the noise, improve signal and the ratio of signal noise so that the speed, vibration, frequency, place can help explainers to analyze and explain the image information. Some common methods used are: digital wave filtering technology, reverse, and wave equation reflection, which in the end will strengthen the processing of data information and improve the quality of image in order that it can be recognized.The ground mine detection radar holds the features of nonstability, non-linear attenuation, the good frequency locality of the small wave transferrance. Therefore, by extension and contraction of measurement, the features of the size of time frequency detection rate can be applied to process the the echo wave non stabilized signal. The application of Mallat calculation based on the small wave high detection rate theory, can deconstruct and reconstruct the land mine echo wave signal and realize the efficient denoise of signal. The experiment in this paper proves that by utilizing the small wave threshold denoise method, the scanning radar A scan denoise process of echo wave signal can be realized and by utilizing the small wave deconstruction and reconstruction sanning radar B scan echo wave processing can be processed, and the low frequency part can be effectively distilled.Through comparing and contrasting the image features of other smaller targets (such as metal fragments, bricks), and basing on the unique tri-hyperbola features of the mine detection image, this paper presents the mine target recognition principle and method based on the vector machine. In testing the correctness andapplicability of this principle and method, the calculation method of such mine target detection is given. The experiment analyzes the influence of the change of the main variables on the mine detection and provides the range of the main variables. The result of the experiment shows that the approach and calculation method of mine target detection based on the supporting vector machine is effective and applicable.