Target Image Acquisition. Strong Background Light In The Principles Of System Design

Recognition of a target image from strong background light has found wide application in target recognition, robort vision, image tracking and industrial inspection. Usually, target image extraction from strong background light employed a series of traditional morphology methods consisting of neural network and wavelet transform. The neural network algorithm needs to introduce complicated mathematical model, has disadvantages of massive numerical calculation, high cost, long exploiting period and is only valid only in specific environment. In strong background, most of the methods based on digital image process from strong background light interference in industrial spo will present possible error judgement and extract false target image.A scheme combining of optical technique and signal process is proposed in the thesis, which uses the monochromatic laser modulated by a signal to illuminate the concerned region. The images of the target are synchronically taken by a CCD. This scheme employs optical narrow filter and discriminative thchniques, which has been used in quite different fields. In the image discriminative thchnique, the grey signal of image are multiplied by the reference signal firstly, and then integrated. The final image is got by digital image processing. The experimental results in sun and bulb light background demonstrate the powerful capability to restrain the background light. Even the target image is concealed by the background light, this technique is capable of recognizing targets in a concerned area illuminated by the laser.The experimental sets up a D/A output board installing in PCI bus of computer to output simulate sine signal to modulate semiconductor laser and a CCD image sensor connected to USB interface to acquire image synchronically; Uses digital image processing algorithm in the computer to realizes the target image enhancement.Uses the object-oriented technology, realizes encapsulation of modulation signal output module, image gathering module, data computation module as well as digital image processing module, and designs the interface function, to call in software. In Windows XP operating environment, in use of operating system's multi-thread technology, develops the software system.This thesis indroduce the optical filter and phase discriminative technique into the image reconnition field. The experiment result has a great significance for further work.