Research On Image Processing Of Vehicle Bench Vibration Test

At present, the vehicle bench vibration test-bed is uses as the main experiment method and platform in examining the vehicles each part's performance. Through gathering the images of the different situation of vibration of the vehicles surface's displacement and the speed signals change's, can let us understand the change situation of the vehicles shell's stress, the rigidity, and the intensity overall . This has certain guiding sense to the vehicles shell's design and the improvement. This thesis is mainly to uses the vehicles shell different position image which CCD gathers to carry on a series of processing, including the image compression processing, transformation processing, pseudo-color processing as well as to panorama splicing processing and so on. Combining the CCD images with the laser vibrating system images, which have the same position displacement images or the stress images, can realize the complete vehicle's visual display which is including stress distribution, the displacement distribution as well as the intensity distribution.The thesis adopted Huffman coding to compress the images. This coding method is through binary tree to representative data. Huffman coding is to represent data by codes of variable length, with more frequent data being represented by shorter codes, with less frequent data being represented by longer codes. While improving efficiency of compressing, reducing the image pixel redundancy, and achieving the image lossless compression.Image pseudo-color processing can transform the gray image to the color image, improving the picture's detail discrimination, while achieving image enhancement purpose. Because any spectrum color can use the three primary colors to obtain according to different proportion mix, when R, G, B have different value, may produce the blending color.According to the references, giving one kind of special pseudo-color processing method—rainbow-coding. The Fourier transform is a kind of common orthogonal transform. We can through two-dimensional Fourier transform of the image to obtain the high frequency which representative slight part—edge and noise, the low frequency representative cheap part, namely smooth region of the whole gray image—the hazy picture.Image mosaic is the art of constructing scene representation with wide view field by use of image sequences shot from real world. It is a kind of problem in image splicing processing that the photos taken with a hand-held camera usually have large perspective distortion, small overlapping, small concentric errors and camera rotation. For it, the classical method of phase correlation is investigated in the thesis and has realized automatic sequence image mosaic. Finally made some simple theories introduce to panorama mosaic, realized the image cylinder projection and panorama splicing of vehicle images.