Processing of slow motion images out of underwater, ground and aerial vehicles

Image processing technology has the potential of producing and reestablishing an actual array of pixels for representing objects to enhance raw images from direct measurements. Some basic requirements consist of both linear and nonlinear noise smoothing, contrast improvement of image brightness, elimination of blurring arisen from atmospheric disturbances (temperature, pressure, shading, etc). When image acquisition is subject to movement between cameras and objects, related geometrical distortions need to be corrected as well. The investigated sample images are captured from real world moving vehicles, namely, Hawaii ocean floor´s image from a submarine, California tropical plants´ image from a passenger car, volcanic Mount St. Helen´s image from an airplane. Both environmental disturbances and slow motions result in the blurring and distortion of real object observations. Taking the distance factor into account, these pictures may be regarded as slow motion images. On the other hand, there are unavoidably various types of noises, like white noise, thermal noise, amplifier noise and quantization noise. These noises are roughly classified into slowly varying noise and rapidly varying noise. Linear approach can be simply applied to noise smoothing of the former case and a nonlinear approach must be selected for the latter case. Image contrast is enhanced using adaptive neighborhood image processing prior to the image reconstruction procedure. A thorough image processing and restoration procedure for slow motion underwater, ground and aerial images is then formulated. In a preliminary case study, comparisons are made between the raw images and the restored intrinsic images, where satisfactory results have been obtained. This complete procedure could be extended to the real time image processing