Text recognition and extraction from videos suffers out of several issues. We introduce a hybrid method for edge detection in this study. We use the product of sobel and canny which proves to be effective when compared to the other edge detection mechanisms. Sobel, Canny Product (SCP) intensify the text pixels and helps in better understanding of the text in video. Once after detecting edges text classification is made that which eliminates the unwanted non text from the frame. This classification is made effective with the help of connected component method. Thus true text pixels are formed out of this classification technique. The resulting frame is then intersected with the Laplacian to obtain the text candidates. Bounding Box formation is the next stage in which text traversal is made. This method flip through the frame from top to bottom and forms a fringe over the text present in the frame. False positive elimination is the final stage, which aids in getting the true text pixels. According to the experimental results, the suggested system performs more stimulatingly than the current approaches.