Shipping Automatic Recognition System Based on Object Feature Extraction

ABSTRACT

During the study on shipping object recognition, we propose a scheme based on improved canny also extract the moving object in video monitoring. The lifting wavelet transformation is adopted to provide smoothing detection for the image edges and the smoothed images are analyzed and evaluated in this paper. We also use improved Gauss function for traditional edge detection to smooth cubic b-spline lifting wavelet. It adopts continuous frame difference method to process the motivating area and obtains edge information by Canny detection. Then an improved Hu invariant moment algorithm is proposed to extract the invariant moment features of object images, which is also used to make automatic recognition for moving objects. The experimental results demonstrate the invariant moment values of the samples, acquired by Canny operator splitting, have obvious clustering effect. So it can offer more accurate recognition for shipping images.

Keywords: