The KLT Feature Tracking Algorithm in Embedded Hardware project is focused to execute the Kanade-Lucas-Tomasi feature tracking Algorithm under embedded hardware to get to know of the mice. KLT algorithm is modeled to choose the best characters and track from one picture to another. This could be optimized to track the characters over modifying the combination of parameter digits. Selecting feature (known as Description) in Picture Processing is about to get the variables which has the output in a few of the interested quantitative data. Feature selection is important to change the raw picture pixel information to the particular method for PC execution.

 We make use of the KLT tracking algorithm to track the mice and to choose the eyes like the two characters to track. To track the mice, the hardware execution provides the benefits to have the wireless “minimum bandwidth” connected to the host executing PC. It removes the requirement to interact with the great quantity of raw pixel information which needs to save further execution in PC. KLT operates including raw picture information in the absence of pictures to be in a particular method. The movement of the mice is taken and the raw information including the dimension of the picture is fetched in the embedded hardware that includes the mice eyes and produce the output like (x,y) co-ordinates of a two dimensional picture.

The hardware utilized for this execution is TMS320C5510 DSP starter kit. The Code Composer Studio™ DSK improvement medium is used for the Integrated Development Environment (IDE). The great attempt under this review is to comprehend the KLT algorithm and to know about DSK kit for productive execution.


The KLT Feature Tracking Algorithm in Embedded Hardware  project is planned on four levels. They are to study the KLT theory, KLT implementation details, familiarize with the TMS320C5510 kit, and execute the algorithm in the DSK kit.

Download KLT Feature Tracking Algorithm in Embedded Hardware