On Target Detection Of Aerial Image Fusion Technology

Since the successful earth observation was realized by the first Synthetic Aperture Radar Satellite"Seasat"in 1978, the technique of Synthetic Aperture Radar (SAR) had developed by leaps and bounds. By far, America, Japan and some countries in Europe throw much manpower and material recourses in succession into this field.The basic law of SAR is that when the satellite flies in orbit, it sends electromagnetic pulse from different places at regular intervals to the same object on Earth, meanwhile, receives echo signals. In a sense, it is believed that this technique extends the length of radar and greatly improves resolution. In addition, SAR can be widely applied in civil and military areas for its exclusiveness of influence of all-day and all-weather atmospheric dissemination and climate as well as its advantage of strong penetration.Target detection is one of the functions of modern radar. For military, target detection and location of SAR possess its vital significance, which are the problems that SAR must solve. As a part of tactical investigation, SAR need observe war field timely and clearly in order to provide more information for evaluation, command and control of the situation. As there are a lot of targets in the war field, such as naval vessels on the sea, cars, trains and tanks on earth, the observation of these targets is one of the ways to get the information about enemy military purpose. This manoeuvre of military institutions may mean the new military strategy of enemy, to which should be attached great importance. What's more, the targets in the war field may have direct danger, such as military vehicles, mobile rocket launchers and self-propelled artillery, which may threat the radar platform. Therefore, the watch on targets on earth and imaging become the key point of the research of observation of the civil airborne and spaceborne on radar on earth. For the application of SAR in civil area, the watch on targets has the same significance. For example, the observation on the motion of vehicles on the ground and the operation of ships on the sea can provide the necessary control information for the management of communication.Based on the technique of image segmentation and target recognition of SAR, in light of the complex background clutter and noises, this paper aims at making a comprehensive analysis, putting forward target detection technique by applying the image fusion technique, which makes the wavelet resolution fusion between SAR image and visual image firstly, makes two kinds of sensors complement each other and is of benefit to object extraction. The targets in this paper are as follows: Firstly, this paper makes a study of imaging principle of SAR sensors and image characters and then puts forth image pretreatment method of SAR based on image segmentation and morphology processing. This paper also makes an analysis of noises removal method of imaging of SAR, and then brings forth the improved noises removal method of wavelet soft-threshold, at the same time, discusses the advantage and disadvantage of several common noises removal methods, such as the comparison of methods of median filter, wiener filter and prewitt with that of wavelet soft-threshold. This paper comprehensively applies all kinds of classic methods and makes edge detection of SAR image and visible image respectively.Secondly, this paper has discussion about the common registration method as well as the registration method used in SAR image and visible image. It summarizes the present image registration methods and with the edge detection technique, put forward the improved selecting method of registration control points which can make registration between SAR image and visible image, meanwhile, on the selecting of the control points this technique uses edge angle point, which can reduce the difficulties of selecting edge angle point caused by coherent noise and make the control point more precise. At the stage of feature matching, the method of selecting feature point by eyes instead of by registration meets the precise, real-time need of intelligence and reconnaissance, and the affine transformation and the bilinear interpolation realize the registration between SAR image and visible image.Thirdly, the paper discusses multi-resolution characteristics of image and elaborates the realization of the process of multi-resolution and signal as well as the process of wavelet multiresolution image decomposition and reconstruction. It mainly talks over fusion algorithm based on SAR image and wavelet transform domain of optical image and analyzes the basic law of wavelet fusion, then definites the method of realization of wavelet fusion. The law says that first, SAR image and visible image are to be decomposed, then the images after being decomposed are to be fused and transformed, finally the images are to be analyzed in order to testify the effectiveness of fusion.Fourthly, this paper uses wavelet energy to detect the targets of fused image, extracts the object by adaptive thresholding and compares it with that of the original SAR image so as to prove the effectiveness of fusion technique used in target detection. As the outline of fused targets becomes clearer and clearer, the information available greatly increases and the images of the two sensors complement each other, it is firmly believed that with the development of the technique of SAR image, the image fusion technique will be widely applied in target detection.