Evaluation and denoising of high-speed image noise based on power spectrum and wavelet denoising

Using high-speed visual equipment is an effective method to locate mobile targets. Under the circumstance of high sensitivity(500Hz), except for the Gaussian noise, atmospheric instability has also an important impact on the image quality. To solve the problem, a method is proposed in this paper based on image power spectrum to analyze and evaluate the Gaussian noise, atmospheric noise meanwhile combined with wavelet denoising to remove the noise aiming at the images acquired by DALSA's 8192*32 high-sensitivity camera. Firstly, image databases are established based on the outdoor working conditions, including normal images, Gaussian noise images loaded with different simulated characteristics and atmospheric noise images in different simulated frequencies. Power spectrum ratio of all the images in the databases is calculated, and the image power spectrum critical value is determined. Then the evaluation and classification of the image noise is got according to the databases and the threshold. wavelet denoising is introduced to remove the noise subsequently. Finally, the comparison of power spectrum between the image untreated and treated is made to evaluate the effect of the method above. Experimental results show that the way can evaluate and remove the noise of image effectively for high-speed visual images.