< Projeler Sayfası

Proje Adı
Geniş Ölçekte Video İşleme

Şirket Adı
Karel

Şirket Mentoru
Dr. Enis Ungan

Akademik Mentor
Yrd. Doç. Tolga Çukur

Asistan
Necip Kurt

Takım Üyeleri
Alperen Kibar, Berk Turhan, Berat Osmanbeyoğlu, Volkan Toklu, Orhun Tomruk

Summary
The aim of this project is to provide a basis in processing high definition video signals through 2D filtering operations such as edge detection, sharpening or blurring on FPGA. The user will be able to configure the FPGA by giving different filter coefficients and observe the output. We have a 7x7 generic filter and we perform the any filtering operation with respect to the input video and given any filter coefficients through this generic structure. We handle taking the input video and convolving it in 1 clock cycle therefore, we give the desired output video. Since we have a limited space in the FPGA we performed our convolution algorithm with the usage of Mega Wizard function that exists in the tools of the Quartus II, so that we achieved the perform the operation in a limited space. Besides, in order to get a proper normalization process which is crucial in our project because of the quantization problem for each pixel value in image or video, we used the algorithm that selects the relevant 8 bits. Lastly, accomplishments according to content of the project can be listed as: implementing 2 dimensional convolution process on FPGA, creating a generic structure for filtering (which has maximum size of 7x7) and filtering process on run time which may be applied to real case scenarios like video surveillance or movie industry.