Air-Conditioning Temperature Regulator Based On Intelligent Person Counter

For the next quality and comfortable lifestyle, many people consider air conditioners (acs) way more than before. As of now, most people have experienced an instance where there's a need to change the temperature settings of the AC for them to be comfortable. To automize this process, this study developed a system that adjusts the temperature of an air-conditioning unit automatically based on the interval of the number of people inside the room set by the user. It uses a 5MP webcam installed overhead the entryway to count the number of people going in and out of the room. The input videos undergo video processing including frame resizing and image conversion before counting the person. To increase accuracy, a pre-trained CNN model is used for people detection. Utilizing python as its language, it'll send the data on the Adruino and commands to the AC through an IR transmitter comparable to the AC's original remote. The system adjusts the AC temperature based on the person's counts inside a room. Results showed that the average percentage accuracy was 85.242% with a percentage error of 14.758%. Slight errors were due to the usage of low-spec camera, a 5MP webcam which decreases the resolution capacity. It was found out the camera resolution was one of the main effectors of the performance of the application counter. All in all, this study works efficiently but it doesn't ensure 100% accuracy. Nevertheless, this study can use in real-world applications.