For the monitoring of indoor behaviors and abnormal behaviors of the elderly living alone, a home behavior monitoring system based on orientation perception is developed and applied in real life. First, this system takes the rudder as the origin, establishes the indoor coordinate system, and divides the indoor area into different interest areas. Then, it connects the Raspberry Pi, camera, and rudder by WiFi, and Python programs are developed to execute the automatic target search of the camera, face recognition, and target location. In the meanwhile, the rudder position, video, and the residence time of the elderly in hot spots are recorded according to the results of face recognition. Finally, on the basis of the above orientation and time information, we analyze the indoor behaviors of elderly people, make abnormal behavior judgment, and send an e-mail alert. The real-test results show that the monitoring system can collect video and orientation information of a target in real time, analyze the target behavior status, and send an e-mail alert in view of the abnormal behavior of the target. In addition, the system is easy to implement and provides rich historical data. Hence, it has practical application value and good application prospect.