|
The underwater laser imaging detection is an effective method of detecting short distance target underwater as an
important complement of sonar detection. With the development of underwater laser imaging technology and underwater
vehicle technology, the underwater automatic target identification has gotten more and more attention, and is a research
difficulty in the area of underwater optical imaging information processing. Today, underwater automatic target
identification based on optical imaging is usually realized with the method of digital circuit software programming. The
algorithm realization and control of this method is very flexible. However, the optical imaging information is 2D image
even 3D image, the amount of imaging processing information is abundant, so the electronic hardware with pure digital
algorithm will need long identification time and is hard to meet the demands of real-time identification. If adopt
computer parallel processing, the identification speed can be improved, but it will increase complexity, size and power
consumption. This paper attempts to apply optical correlation identification technology to realize underwater automatic
target identification. The optics correlation identification technology utilizes the Fourier transform characteristic of
Fourier lens which can accomplish Fourier transform of image information in the level of nanosecond, and optical space
interconnection calculation has the features of parallel, high speed, large capacity and high resolution, combines the
flexibility of calculation and control of digital circuit method to realize optoelectronic hybrid identification mode. We
reduce theoretical formulation of correlation identification and analyze the principle of optical correlation identification,
and write MATLAB simulation program. We adopt single frame image obtained in underwater range gating laser
imaging to identify, and through identifying and locating the different positions of target, we can improve the speed and
orientation efficiency of target identification effectively, and validate the feasibility of this method primarily.
|
