Intelligent transportation is expected to be a part of smart cities in the near future. Vehicles equipped with multiple cameras that convey traffic information to the driver would be an important part of transportation. In this paper, we consider the problem of hardware-efficient stitching of videos captured by multiple cameras. An approach based on detection of interest points in all the video streams to be stitched is proposed. A VLSI architecture for interest point detection is developed. An architecture for display of two video streams is also presented. The proposed architecture has been implemented on an FPGA and data on device utilization is presented. © 2015 IEEE.