Research On Laser Active Imaging And Recognition Technology

Laser active imaging technology has been developing fast in the region of activeimaging and detection in recent years. It not only overcomes many shortcomings ofpassive imaging technology, but detects object in the whole day, and has manyadvantages including clear image, high contrast, good anti-interference. It becomes animportant application technology of photoelectron detection and alarm equipmentbecause of its own distinctive advatages.To prevent terrorists carrying dangerous goods invading and stealing preciousproperties in import position, such as bank, heritage, treasury, warehouses, oil factoryamd so on, to enhance combat effectiveness of counter-terrorism, to accelerate theability to obtain detection information, and to improve automation degree of detectionsystem, a new kind of laser active imaging and recognition system were developedcombining image recognition technology with classical laser active imaging system,which can obtain high resolution information of detection object, and make intelligentautomation recognition for obtained image information.This paper carries out research with laser active imaging and recognitiontechnology. Fistly, laser propagation in atmospheric environment, model of activeimaging range and mode of signal to noise ratio was introduced and analyzed,respectively, and we know that imaging range and signal to noise of system can beincreased by increasing modulation transfer function of system or laser power,decreasing F number of lens or laser divergence angle, improve optical transmittanceof laser emission system and image receiving system, which take guidance for systemdesign of laser emission and image receiving; Then, active imaging system design andobject recognition technology was researched, designed system scheme make enoughpreparation for developing equipment and building experiment platform; And then,we built experiment platform of laser active imaging and recognition system byLU0793T030fiber coupled laser diode, Falcon HG1M120CMOS camera, YS3081bearing turntable and developed control operation software interface. We obtained recognition probability of68.87%and72.11%with designed experiment procedureand BP neural network algorithm for detection objection——submachine gun modelat450m and550m from experiment system at night, respectively. In the end, we putforward further improvement to experiment results and further upgrade to experimentplatform.