| Currently, aging has become a worldwide problem all over the world. According to the forecasts, China will become the highest degree of aging of the country in2030, and even to reach one third when2050. What is even worse, the social security system is not perfect, pension issues in becoming a crisis, a crisis in which the most important point is that the elderly living alone casualties. On the other hand, in the field of computer vi-sion, pattern recognition-based rapid development, making the image recognition tech-nology used in everyday life possible. Thus, people with a computer instead of caring for the elderly, reduce the occurrence of unforeseen circumstances to be an effective so-lution. Therefore, this paper will examine the use of image recognition technology, in recognition on the basis of body posture, build a home for elderly people living alone monitoring system. Thus the main work of this paper has the following aspects:First, target detection. Analysis of different target detection algorithm, and considering environmental factors, using the background model based on Codebook model elimi-nates target detection method, and by using the authentication method used is reasona-ble.Second feature extraction gesture. Feature extraction methods to compare different attitude, and choose to convert the detected target star skeleton model, extracted from human gesture features. Posture while using Codebook model features a preliminary judgment.Third, the motion gesture recognition. Noting attitude is part of any movement, from the movement of gesture is pointless. The use of human motion with Markov property, combined with hidden Markov model to complete the identification of the current posture, and through experimental testing algorithms.And the last, realizing a monitoring system. On the basis of the above study algo-rithm based on the use of cross-platform programming environment designed to achieve a basic intelligent monitoring systems. |
PDF Full Text Request