Research On Demand Adaptive Air Supply Of Air Conditioning Energy Saving Base On Computer Vision

In resent years, with the rapid development of economy and civil construction in China, the energy consumption in the buildings becomes greater and greater in the total consumption. The problem of electricity supply emergency exists in many cities in China. The air conditioning system is the main consumer of electric power and its energy consumption problem is becoming increasingly conspicuous. Therefore, in the view of the growing tension in energy supply and the people's high requirement of the indoor environment and indoor air quality, how to maintain the air conditioning regional comforts, which also under the premise of minimizing the energy consumption of air conditioning, is the current problem that needs to resolve urgently.The paper studies the control of air conditioning energy saving based on the actual need, which is a control method of air conditioning energy saving that based on the relationship of the "supply and demand", which adds fairly important values both theoretically and practically to this field. The paper hereon makes an introduction of omni directional vision sensor (ODVS) as detection technique, and uses the ODVS to detect the indoor "need", which represents the actual num of indoor people. And according to the actual "need", the paper implements the real-time duty cycle control, which tries to solve the current control problems of uncomfortableness, slow responding speed and energy inefficiency.The paper firstly analysis the air conditioning energy saving potential at present and explores how to supply on demand in the aspects of comfortableness and energy saving, and puts forward the detect and control methods of excavating the potential of air conditioning energy saving from two views of satisfying maximal comfortableness and healthy situation. Then, based on the two different air conditioning energy objects, the paper presents the different energy saving control methods; central air conditioning energy saving control method directly adjusts gas source capacity and regulates the size of fresh air valve based on the actual detection of the "need"; ordinary high power air conditioning uses the duty cycle control technology to achieve the "need" adaptive control. The duty control without affecting the comfortableness of indoor people is a technology that the maximum use of the air conditioning energy saving potential. Although different air conditioning energy saving object has different control method, the core problem of the air conditioning energy saving is how to acquire the number of indoor people, so from the methods of image acquisition and image processing, the paper focuses on to introduce how to detect the "need".Against the means of detect the "need", the paper establishes the fixed background model and the adaptive Gaussian mixture distribution background model to achieve the static and the dynamic movement target detection. Then, it designs the human body rectangular model that based on the characteristics of the ODVS, and judges the human area and shape attributes, which also combines with the projection analysis and the skin color detection algorithm to acquire the number of indoor people. At the same time, the real-time duty cycle control curve is given in accordance with the number of people. Finally, the paper develops a testing energy saving control of the air conditioning system based on the Java language and introduces the steps of modules in detail and tests in different environment. Experiments show the system has the small amount of calculation, the strong real-time performance and the fairly good robustness.