Making use the prior knowledge that higher order cumulant and wavlet transformation methods are
combined, which is used in the multi-dim-small target detection with the Gaussian noise and natural
background. In this paper, a method of magnifying image based on two dimension discrete wavelet
is introduced. Comparing with traditional interpolation methods, it not only enhanced resolution of
the aero-image, but maintained the detail character of the image. Experiment results have shown
that it is good method for enhancing aero-image. The same aerial image is magnified with
above-mentioned wavelet-based algorithms, conventional bilinear interpolation and tri-spline sampling
interpolation algorithm separately. The values of mean gradient on three graphs are calculated using
three approach, they are 6.31, 7.43 and 8.45. In accordance with the result, the value of the magnified
image by this method is largest.
Proc. SPIE. 7658, 5th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optoelectronic Materials and Devices for Detector, Imager, Display, and Energy Conversion Technology
KEYWORDS: Target detection, Image fusion, Detection and tracking algorithms, Image segmentation, Digital filtering, Wavelets, Electronic filtering, Filtering (signal processing), Fusion energy, Data fusion
A novel two-dimensional (2-D) adaptive filtering algorithm to enhance 2-D signals of small spatial extent embedded in
white Gaussian noise and a variety of outstanding modern digital signal processing methods are combined in background
segmentation. The coefficients of the adaptive filter converge to a special 2-D slice of the fourth-order cumulant function
of the input signal. The proposed algorithm is called the 2-D cumulant adaptive enhancer (2DCAE). And the methods in
background segmentation is combining the wavelet energy transformation, data integration and threshold segmentation.
Targets tracking methods based digital image are used widely. During the process of multi-targets tracking, when the
targets falling in the tracking door are not only one, joint data association is used as a traditional method. But there is so
much operation in this method that the need of real time can not be meet. Targets tracking relating method based on
fuzzy reasoning and expending Kalman filter is adopted in this paper.
Firstly, wavelet energy is adopted to detect targets in the image. The wavelet of bior1.3 is used to decompose image.
It is anti-symmetric wavelet and is fit for detecting variational information in horizontal and vertical direction. Because
the detected image generally have natural texture background, such as sea, sky, and so on. The targets have high
frequency characteristic at horizontal and vertical direction in these images, but the background has high frequency
characteristic at one of the two direction. The wavelet energy of targets at two direction go beyond that of background.
Then through selecting self-adaptive threshold, targets and background can be divided. By the method of figure centers
distance classifying, the figure centers of every target can be calculated. Secondly, Targets figure-center coordinate is
detecting through self-adapting figure-center filter.
Thirdly, fuzzy reasoning method is used in the progress of data relating. It is easy in engineering application
compared with the JPDA arithmetic. In the experiment, after several frames relating, the mean square error of row
coordinate reduce from 0.278 to 0.148. The errors of multi-targets figure-center coordinate gradually decrease. The
forecast veracity gradually increase.
The technology of multi-target tracking is always detecting small targets in complex background. According to the frequency characteristic of nature background and small target, a method of detecting targets based on wavelet energy is suggested. The energy of targets in horizontal and vertical direction outclass that of background, or we can say, the gray degree of targets outclass that of background. Then through selecting adaptive region-value, targets and background can be divided. By the method of classifying, the figure centers of every target can be calculated. According to the figure centers of every target in several neighboring frames, the position and velocity of every target in next frame can be estimated by Kalman filter. Experiment results have shown that small targets in sea or sky nature background can be detected by this method, and they can be tracking.
A wavelet-based approach of aerial blurred image restoration is proposed in this article. Image motion is inevitable in photographing for aerospace camera. Though some Image Motion Compensation(IMC) schemes are applied in aerial imaging system, the ultimate image will be blurred in certain extent for the existence of IMC residual error, while the forward image motion is the key element among all image motions which lead to image blurring. First the course of blurring caused by forward image motion is expressed using wavelet transform, and a multiresolution sparse matrix representation of the degeneration model is obtained according to the wavelet transform. Subsequently a regularizing restoration algorithm is deduced from it, and which can smoothly restraint the processed result efficiently. In the end the proposed approach is tested in MATLAB. The blurred image is restored using above-mentioned wavelet algorithm, conventional contrary filter and Wiener filter algorithm separately. The conclusion that the wavelet-based restoration algorithm is superior to other two approaches is obtained by comparing the restored image’s value of mean gradient. The calculating quantity of the wavelet-based blurred image restoration approach isn't large and it has good practicability in the field of image interpretation and aerial survey or drawing.
The resolution of aerial digital image is low because of atmosphere's swaying and the Image Motion Compensation's error while imaging. In view of the limitation of ordinary interpolation an approach of image magnifying based on two-dimension discrete wavelet is proposed in this article. At beginning the original image is magnified using tri-spline sampling interpolation. Then we perform two-dimension discrete wavelet transformation to the magnified image. Subsequently the inverse transformation of wavelet to recombine digital image is carried out utilizing the acquired three high-frequency images as high-frequency part of the inverse transformation and utilizing the original image as its low-frequency part, and the obtained image is double to original image in the direction of length and width. In the end we completed the proposed approach in MATLAB. The same aerial image is magnified with above-mentioned wavelet-based algorithms, conventional bilinear interpolation and tri-spline sampling interpolation algorithm separately. After comparing the magnified image's value of mean gradient, the conclusion is deduced that the wavelet-based approach can not only magnify aerial image and enhance its resolution better, but also retain the minutia of original image, through which the aerial image is ready for the further aerial image interpretation and processing.