Date of Award


Document Type

Open Access Master's Report

Degree Name

Master of Science in Electrical Engineering (MS)

Administrative Home Department

Department of Electrical and Computer Engineering

Advisor 1

Seyed (Reza) Zekavat

Committee Member 1

Michael Roggemann

Committee Member 2

Mahdi Shahbakhti


Camera sensors are emerging in many applications such as Smart Buildings and autonomous driving. The Data generated by multiple cameras in a smart building and autonomous driving applications is usually transmitted through an edge box to a cloud terminal. This transmitted information requires a considerable channel bandwidth, which is not available through current communication standards. The report proposes a Camera Sensor Frame Reduction method to decrease the required channel bandwidth for applications such as autonomous driving.

Here, we propose a method that incorporates cross frame similarity measurement method to reduce the redundant frames and decrease the data rate of each camera. This approach adds processing to the camera sensor, which maps each camera to a smart one. In order to calculate cross frame correlation, each smart camera converts frames into blocks of sub-images. Next, we incorporate consecutive blocks to compute the overall cross frame correlation. The report studies block size selection and its impact on processing complexity and performance. We used real vehicle videos in different driving speed and scenarios to study the complexity and performance of the proposed method. We have investigated frame reduction rate as a function of vehicle traffic and driving environment.