Previously, D.W. Tindall (1995) discussed some of the challengesassociated with automatic licence plate recognition (ALPR) and technicalsolutions based on digital signal processing, including the use ofneural network recognition algorithms. Since that time, Racal hasinstalled and commissioned systems worth over £1M and has gainedconsiderable practical experience. Contracts from overseas are now alsobeing awarded. In nearly every instance, the customer requiresrecognition of licence plates from countries other than the nativeformat, and this places further challenges upon the DSP algorithms andthe available processing power. Racal has continued to develop the ALPRproduct family and associated subsystems, and the paper presents some ofthe experiences gained so far and how we believe ALPR systems willdevelop in the next few years. The first part of the paper concentrateson the variability of licence plates in a multinational environment andsolutions to some of these challenges. The second part discusses thedevelopment of practical systems which offer customers an effective andautomatic means for monitoring traffic in both controlled and open roadscenarios
展开▼
