The Senegalese highway toll collection system is inefficient, and significant congestion and delays frequently occur at toll booths in the face of sudden traffic surges. This paper examines the limitations of the existing toll collection methods on Senegalese highways and combines the image detection algorithm and electronic license plate recognition technology with the existing MTC and ETC toll collection methods to construct a fusion toll collection method based on vehicle information recognition. The BPV Thiès toll station is selected as an example, and experimental simulation is carried out for different deployment schemes. Compared with the status quo toll collection method, the inbound delay can be reduced by 34.99% and 62.49%, and the outbound delay can be reduced by 38.01% and 62.44%, respectively, during peak hours. Upon increasing the ETC penetration rate to 50%, the convergent tolling scheme further reduces the average delays at inbound and outbound stations by 80.57% and 79.64%, respectively. This demonstrates that the fusion tolling technique can markedly enhance toll collection efficiency and serves as a pragmatic reference for highway toll collection methods in Senegal.