Abstarct: Nowadays, digital video files which contain voice and image data are accessible anywhere. Rapid development of digital video equipment such as digital cameras has extremely widened application area of video including video conferencing and electronic education. Therefore, a large amount of video data has been created which the need to handling and organizing such a large databaxse has caused shot boundary detection to be an active research area in digital video processing field. In this thesis, we aim to design and implement a video shot change and shot type (i.e. gradual and abrupt change) detection method such that is real-time implementable on DSP processors platform. We perform this DSP implementation in both software (using Code Composer Studio®) and hardware (using an available Evaluation Module-EVM). Our proposed method consumes low memory and has high execution speed. We implemented this method on an EVM baxsed on TMS320C5505 DSP processor. In simulation, we evaluated the proposed method efficiency using precision, recall, and mean run time measures for each frxame of our collected video sequences. This video collection consists of four categories of sport, news, cinema, and cartoon. We reached 95.2% and 94.5% for precision and recall measures respectively and 0.095 seconds for mean run time measure. For DSP implementation, we computed and reported the computational complexity in terms of consumed machine cycle and total number of multiplication and summation. After implementation on TMS320C5505 DSP platform, we tried to real-time implement our proposed method on TMS320DM6446 digital video processor which is still continuing, although it is difficult because of not access to useful resources and professional individuals. In this thesis, we report the respective results obtained till now. We gathered a relatively large video databaxse using a digital/analog TV signal capture with specified features such as frxame rate and content type.