Method for recognizing image scale invariant pattern under noise condition

Abstract

The invention discloses a method for recognizing an image scale invariant pattern under the noise condition. The method comprises two separated stages: a training stage of training the image of a specific object, namely a reference image, to obtain a correlation filter, and a correlation detection stage of recognizing the object to be detected by using the obtained correlation filter. In the method, noise estimation can change adaptively with the change of the signal to noise ratio, and high noise immunity is achieved. The main principle of the method is that: the image can be reconstructed by an intrinsic mode function correlated to the important structure of the image, thus the correlation filter can be built by some specific layers (the intrinsic mode function) to realize the scale invariant pattern recognition; moreover, as the noise distribution on the intrinsic mode function of each layer has a certain rule, the aim of eliminating noise can be fulfilled by the noise distributionrule.

Description

Technical field [0001] The invention relates to an image processing method. In particular, it is a method that uses a computer to detect and recognize objects of different scales in images under noisy conditions. Background technique [0002] One of the main purposes of pattern recognition is to study methods that can perform distortion-invariant recognition (for example, detection methods that are not sensitive to input object distortion). In image processing, three typical object distortions are translation, rotation and scale change. A lot of research has been done on these three kinds of distortions, and some methods have been proposed to obtain distortion invariant detection. Most of these methods are based on the idea of ​​correlation filters. The translation of the object in the image can be easily detected by the Fourier spectrum. If the object in the image rotates in the plane, its Fourier spectrum will also produce corresponding rotation and the period is 2π, so the...

AI Technical Summary

Problems solved by technology

However, the correlation peaks output by these methods are not stable due to the scale change factor, and are very sensitive to background noise, and the detection performance of these methods is greatly limited.

The radial harmonic filter based on wavelet transform can remove part of the high-frequency noise, but its biggest disadvantage is that the wavelet basis function must be specified in advance, and different basis functions lead to different detection results

Images