Research Of Infrared Image Target Detection In Ship

As modern science and technology, especially computers and digital image processing technology, the rapid development of infrared thermal imaging technology has become a broad high-tech future. It is widely used in national defense, public security, fire protection, traffic management, defense of key facilities, sea and border control, port supervision, yard warehouse fire early warning and other fields. Infrared thermal imaging and monitoring equipment may work both day and night, and will not expose yourself; Also in rain, fog and other bad weather conditions, the infrared wavelength is longer, through rain, fog, smoke resistant and can be a normal against various targets is monitored. To sum up earlier, it has a hidden good detection ability, the role range, high detection sensitivity, reliability and so on. Therefore infrared thermal imaging system and radar detection system can be used as a supplement in a ship collision avoidance, navigation safety, vessel traffic management to gain the absolute leading edge and an irreplaceable role.As the infrared images of ships in general have a target distance, target shape, weak, details of the missing, low signal to noise ratio and so on. Thus to infrared detection theory and methods of ship research has focused on improving the image signal to noise ratio, detection algorithm and its scope of application, reliability and so on.This thesis mainly deals with the infrared target detection under complex background, infrared images for the object-background characteristics and features of the ship, which is achieved by the methods of analyzed the common algorithm of image pre- processing, image segmentation, and infrared images for the ship and the target signal to noise ratio is low, susceptible to background clutter interference characteristics, concluded the advantage and short coming of different algorithm, then, brought forward a targets detection method. The most important characteristic is: At first a method to classify impulse noise form image, then remove the impulse noise with adaptive Median Filter, in order to reduce the resulting damage to the image detail which was caused by the Median filter algorithm; second, because of when wavelet threshold de-noising method for the de-noising ,the image detail is able to retain maximum. And threshold function and threshold determination is the key of the wavelet threshold de-noising method, in this paper, the new threshold function between the hard threshold and the soft threshold, and improve the traditional method of determining the threshold. So put forward and used improved wavelet threshold de-noising method to process the pixel points which is suffered gauss noise; At last, After the image preprocessing for image quality has been the decline, so extracting the targets for a new Canny edge detection method devised on the basis of an improved fuzzy enhancement .At last, this target detection algorithm was programmed under the matlab, the experiment showed that the infrared ship target detection algorithm the paper suggested is rational in process, viable, and the ratio of detection is high.