Search

When the target is several miles away from the ultra-wide field of view (UWF V) infrared warning system, it will be a point target in the infrared image, so there is no the target information of distance, geometry and texture without which it is hard to assess the threat of target accurately. It is ve ry important for the air defense command and decision making to have a correct threat assessment of the aerial target, and at present there are few reports about the aerial target threat assessment of the UWFV infrar ed warning system. The characteristic of the UWFV infrared image is analyzed. A laser range finder is used to measure the initial distance of each targ et which will be sent back to the infrared warning system. Together with the target information of initial distance, gray value, course angle and angular altitude, considering the nonlinear characteristic of aerial target threat assessment, the threat assessment method based on RBF neural network is presented for its good self-adaptive and self study ability to solve nonlinear complex problems. After simulation experiment, it is found that this method is available and effective. Key words: Ultra-wide field of view (UWFV) infrared image, mu ltiple targets, threat assessment, RBF neural network

An approach of in-situ measuring the complex response of a scene spatial light modulator which working in differential state was presented by using an image of Ronchi grating substituted for a concrete one. Its principle and testing method have been studied and a typical XGA2L11 spatial light modulator has been characterized. The scene image under complicated background has been edge enhanced by mapping the measurement results, and the correlation results showed that scene spatial modulator with the differential state can increase the SNR and PCE of correlation peak. The way could be wholly applied in Vander Lugt Correlator and has very strong practicability and effectiveness.

The detection for motive small target in infrared image sequence has become a hot topic nowadays. Background suppress algorithm based on minim gradient median filter and temporal recursion target detection algorithm are introduced. On the basis of contents previously mentioned, a four stages pipeline structure infrared small target detection process system, which aims at characters of algorithm complexity, large amounts of data to process, high frame frequency and exigent real-time character in this kind of application, is designed and implemented. The logical structure of the system was introduced and the function and signals flows are programmed. The system is composed of two FPGA chips and two DSP chips of TI. According to the function of each part, the system is divided into image preprocess stage, target detection stage, track relation stage and image output stage. The experiment of running algorithms on the system presented in this paper proved that the system could meet acquisition and process of 50Hz 240×320 digital image and the system could real time detect small target with a signal-noise ratio more than 3 reliably. The system achieves the characters of large amount of memory, high real-time processing, excellent extension and favorable interactive interface.

Optical correlator technologies play the important role in image processing systems. However, building a set of optical correlation system in lab, such as Vander Lugt correlator (VLC) or Joint Transform Correlator (JTC), takes lots of time and the expensive price. Meanwhile, there are some key technologies to solve in building processing. As an adjuvant tool, the simulation for optical correlation system by computer becomes an important means in building course. The computer simulation can find out the possibly potential problem in building term and the different operating condition to analyze the different experimental results. The paper offers detail the simulating procedure for JTC and VLC. In simulating term, the joint power spectrum differential method to smooth down the zero-order diffraction spot in JTC and produce an optimizing method to subdivide the macro-pixel based on Lee encoding method in VLC. The comparison between simulating and experimental results proves the feasibility of these methods and shows the bright application field. In the end of paper, an applying sample of simulating software to identify fingerprint images is given.

A method for measuring refractive index of plate glass by joint transform correlative technique is given. In the one of two correlative light beams passing through the two correlative images of certain separation in input plane, a unknown refractive index and known thickness plate glass is inserted with a certainty angle (θ). Joint transform power spectrum (JPS) is recorded separately at the back focal plane of the lens by a CCD detector when Θ = 90° or Θ = 45°, the JPS is sent to the LCD on the input plane by the computer. Then two groups of correlative points will be recorded separately by the CCD detector. On the basis of the interval difference value between two groups of altered correlative points, the refractive index is calculated. Theory analysis and examination result are given.

Multiple synthetic discriminant function (SDF) filters has been verified to be a better strategy for distortion invariant optical pattern recognition, especially for the in-plane rotation invariant ones. However, how to select the training sets is very crucial for the filter’s validity. In this paper, referring to the Rayleigh criteria, the distortion tolerance of a single matched spatial filter (MSF) has been tested by using a digital-optical hybrid 4f correlation system. Experimental results are ±5° and ±8% for the rotation and scaling distortion tolerance, respectively. According to the results, an equal correlation peak phase-only SDF algorithm has been adopted to synthesize two SDF filters, in which the training set images have been correspondingly selected with 5° and 10° intervals. The SDF filter constructed with 10° interval training sets has demonstrated as good performance as that one with 5° interval, but with the advantage of reduction by half of the number of multiple SDF filters for 360° full-rotation, which indicated that the double of distortion tolerance selected as the interval of training images is more suitable for real-time correlator recognitions.

We introduce a multi-point and real-time method for corrects the nonuniformity of uncooled IRFPA. First of all, all pixels are classified to different kinds according to their response characteristics in the system. Then we figure out the correction parameter of each subsection of every kind response curve. Not only the correction precision of this method is higher than two-point correction algorithms, but also it solves the difficult, such as needing lots of correction parameter, the system will be very complex and hard to realize, etc. of common multi-point method. It has marked merits as follow: high correction precision, easy to realize and the real-time characteristic.