Application Research Of Coal Transport Carriage Code Number Based On Vision Recognition

With the improvement of the recognition accuracy of railway Coal transport carriage’s number,the existing RFID identification system has been unable to meet the needs of accuracy and reliability.In order to improve the accuracy and reliability,this paper studied the method of improving the reliability and accuracy of the number recognition of the train carriage in the coal fired power plant based on machine vision.Traditional methods of China’s railway carriage number identification is copied by artificial with the people of Special Rapporteur train section station and marshalling station.This manual way is not only boring,and need high labor intensity,but also is very easy to make mistakes.Later appeared the bar code identification system and the microwave reflection identification system have been eliminated due to the unreliability of technology.Currently used identification system is based on RFID radio frequency technology.However,with the increase of the time,if the environment is serious,such as in the thermal power plant,the railway carriage number identification often present much recognition error,and the recognition accuracy may be unable to meet the needs of production.thus machine vision is a new kind of intelligent recognition technology can be adopted,it has the advantages of non-contact,remote,and obtained much attention in this field.In this paper,the image acquisition system of the dynamic train compartment is designed.The constitution of the system,the selection of the equipment and the control of the system are described.The image acquisition system is composed of industrial camera,photoelectric sensor trigger,supplementary light source and PLC control module.This paper focuses on the image preprocessing and recognition of coal carriage.The weighted average method is taken to convert the RGB image into the gray image,and using the adaptive threshold to make the gray image binaryzation,the image is enhanced by using the method of piecewise linear transformation.Finally,the improved median filtering and morphological opening algorithm are used to remove the salt and pepper noise,then we use mathematical morphology algorithm to locate the character region,using recognition feedback and empirical statistics tosegment three character blocks,and each block is split into a single character,in the last,the BP neural network is used to recognize the segmented characters.Finally,this paper also presents a method to improve the reliability and accuracy of the recognition train coding number for coal transport in the thermal power plant.By adding a computer vision recognition device,combining with the original RFID radio frequency recognition device,a parallel recognition system is established.When the train passes,the carriage images are acquired by camera,and the images are preprocessed,positioned,segmented,recognized by the computer vision system,then the results of RFID and vision system are fused according to the decision rules.Experiment result shows that this method can improve the accuracy and reliability of carriage number recognition.