Summary:

    During my studies, I delved into AI Object Detection, Collection, Storage, and Data Security for my final project of my Python class. ​​My group’s focus was to gather data from fixed traffic cameras and store it for use. The program was to note the time an object was detected, whether it was a car or person, or any other distinguishable objects in the cameras view. The object would only be saved if the model was beyond a certain threshold of perceived accuracy. ​Thereafter the data would be securely uploaded via the RSA-SHA 1 algorithm to the Amazon Web Service Cloud and then sanitized from the uploading computer. The data can be accessed again but only by authorized users.

Protection of Sensitive Data
   Security is essential because it protects sensitive data (like faces and license plates), ensures compliance with legal and privacy regulations, and maintains data integrity through a cryptographic method, RSA-SHA1. This security framework prevents data tampering during transmission, safeguards against man-in-the-middle attacks, and enforces strict access controls so that only authorized users can access the information. Additionally, secure data sanitization minimizes the risk of residual breaches on local systems, building overall system reliability and user trust.

​​​​

Cons:

    This method does not protect the deleted data from being recovered from the hard drive via Data Carving, Magnetic Force Microscopy, or Residual Data Recovery. It also only supports one camera at a time.

​

Post Submission Edits (Solo):

    Incorporated multiple Arduino Web camera’s that host a local webserver and the ability to process more than one feed at once. A Raspberry Pi connects to each web camera and displays its inference and output data on another local webserver. For ease of initial setup, the boards are all wirelessly and automatically disperse Wi-Fi credentials during initial setup from a master board that the user interacts with via a webpage.

​​