Search

In order to improve the efficiency and accuracy of warship target detection in sea-sky background,and according to the obvious gray scale contrast between the sky region and the sea region, this paper operate the division of the two regional context by looking for low-water mark of the histogram's Valley, which is one feature of the histogram called "Two Peaks and a Valley". Then the method obtains the statistics and classification of background pixels, and selects the relevant parameters to establish Gaussian model A and parameter model B. Model A is used to complete the perception of invasion target and to transmit the relevant information to model B immediately,then the model B distinguishes the target with labeling. The experimental results show that the algorithm has a good ability to adapt to the environment, easy to implement and accurate at target detection.

Aim at the problem of sea-sky-line detection in sea-sky background image, a new method which is based on histogram analysis and linear fitting is proposed. It can obtain the information of pixels near the Sea-sky-line through an analysis of the histogram, and calculating the rough area to get rid of irrelevant pixels in order to extract the sea-sky-line through the method of linear fitting. The experimental results show that this method can detect out sea-sky-line under complicated sea-sky background effectively, and has many advantages such as strong robustness, speedy calculation and high practical value.

A de-noising method using wavelet transform and morphology for low light level image is presented. The intensity of noise field is different from the intensity of normal image region, which has reduce the quality of vision and after works of image processing such as segmentation, feature extraction and identification. Since there are a number of noises in low light level image, this paper did some research of finding the source of noise that constitute the main problems and contrasted the existing noise reduction techniques, and then the method proposed makes the best of the multi-resolution characteristic of wavelet transform to process the specific part of image. Experiment indicates that this method can resolve noise reduction problem of reserving image detail and is effective on noise reduction and contrast enhancement in low light level image.