In an aging society, accidental falls are highly likely to occur due to external factors or the incidence of diseases in the elderly. Serious consequences can arise when an elderly person living alone falls and is unable to get up to call for help. To address the issue of automatic assistance for elderly individuals living alone when they fall by accident, and the problem of behavioral analysis in daily life, this paper designs and develops an elderly monitoring system through the study of feature extraction algorithms and posture recognition algorithms. Furthermore, this paper has also developed a set of three-dimensional spatial angle calculation methods suitable for implementation on a microcontroller. The CORDIC algorithm can compute two-dimensional spatial angles through simple addition, subtraction, and bit shifting operations. Through careful study of the CORDIC algorithm, it was found that applying the CORDIC algorithm twice can achieve the calculation of three-dimensional spatial angles. Through experimental testing, we can see that the algorithms proposed in this paper have achieved good experimental results in terms of computational accuracy and operation time, reaching a relatively advanced level.