Once a plate has been recognized, character extraction algorithms might face difficulties stemming from environmental lighting conditions. Low contrast images caused by reflection, low lights at night time and foggy or hazy weather, make it difficult for the OCR algorithm to operate successfully. To address these difficulties, RSIP Vision has constructed an algorithm to extract and identify license plate numbers in low contrast images or when direct sun or synthetic lighting cause reflections.
Accurate number plate binarization is crucial for the correct operation of the OCR module. We therefore put special emphasis on image binarization in our algorithms. A binary image makes it easy to extract edges of individual characters and compare them against a character database for identification, or easily characterize their shape in the case no font prediction is available. RSIP Vision’s algorithm employs a variety of tailor made techniques to binarize the image, overcoming poor lighting condition and reflections and enabling a clear identification of characters on the number plate. Our software can quickly and accurately transmit license plate details to the client’s database for further action. Our software also turns adaptation of the detection procedure to match different formats of number plates into a straightforward task.
It is estimated that by year 2020 the ITS market will attain roughly 50 billion US dollars. Evidently, vision techniques (hardware and software) will play a central part in the core technology of ITS, for either car-mounted or static cameras. As the number of cars on the road increases, and the road system becomes more intricate, automatic monitoring will become crucial. In the last few decades, RSIP Vision has developed a proven record of success in OCR and image processing projects. Please consult our OCR projects page to learn more about how RSIP Vision can help make your ITS project a success.