I know from experience that dogs may not always bark when they want to come in, so I figured I could create a program which would let me know when they are at the door. SO the project was designed to alert the owner that the dog is at the door and wants to come in. This can be inferred as the dog is at the door.

The algorithm is a detection program which uses a camera to detect images, more specifically dogs. The detectnet program used in the algorithm has an already preset dog class id built in, so when used with the camera the program can detect a dog or an object in general and print out what it is. Algorithm works okay in situations where lighting is poor, the camera is not the best, and the dog is lying on his side with wires such as a phone line and a couple of other cables in the background. The results would have probably been better in different circumstances involving better light and a different background.

To run the project you have to download it and make sure that the machine running it has a camera attached to it. You will have to have downloaded jetson inference, jetson utils, argparse, and sys as these are the libraries used to run the project. Once you are ready to run the code use the command sequence python3 name dev/video0. This is assuming that the camera name is zero for the device that is being used. If it is a laptop it may end up being called camera one or another number.