Based On Embedded Video Acquisition Notes Image Distortion Correction

With the development of economy and technology, the embedded image processing is applied more and more widely, has been widely used in military, financial, transportation and other fields. Bill image processing and recognition is bill image recognition technology is an active topic in recent years, the field of pattern recognition, recognition, embedded system of paper currency image combined more closely. Because the imaging equipment of the system is not perfect, the image failed to get to the actual effect, produced out of proportion, and even deformation, this phenomenon is called geometric distortion. Geometric distortion is mainly due to each pixel in the image displacement, specific performance for the image of the object deformation, near far ratio imbalance etc. Geometric distortion is mainly due to each pixel in the image displacement, specific performance for the image of the object deformation, near far ratio imbalance etc.. The distortion of image recognition and the authenticity of banknotes will direct paper money discrimination, will have a serious effect.In this paper, banknote image distortion obtained starting by embedded platform acquisition, briefly analyzes the reasons of geometric distortion produced, in-depth and detailed analysis and Research on paper currency image distortion, analyses the reasons of geometric distortion correction and expounds the principle of digital image geometric rectification, and on this basis, this paper put forward the method of geometric distortion connection point method based on the correction, the method is to set up the connection in the distorted image and standard image, realizing the image restoration by geometric transformation. In the process of distortion correction notes, first, image enhancement, image of currency on the distortion of the histogram equalization enhancement of the image, and then uses the Sobel operator on horizontal and vertical direction edge detection, the edge image obtained corrosion expansion operation, and then through the angle setting on the contour was extracted by Radon transform, obtain banknote image edge location, finally uses the connection point method for correction, the ultimate realization of paper currency image distortion on embedded platform collected correction.