The Design of Intelligent Sensor Network Based on Grid Structure

Brightness obviously affects the pixel value of an image sensor. Image identification from a traditional sensor may generate a bias and surveillance misunderstandings because of weather. The proposed intelligent image sensor integrates an optoelectronic converting circuit and an embedded system and can convert the brightness to an analog voltage. From the voltage value, we can establish a contrast list to show the relationship of the analog voltage and image pixel values. Using the list and applying the linear interpolation method, we can eliminate the effect of the brightness and raise the identification rate. To improve the deficiency of conventional residence surveillance systems in real-time, multipoint, dispersed surveillance, the study proposed an integrated grid technology-based intelligent image surveillance system design; To further improve the congestion phenomenon created by the proposed multipoint, dispersed surveillance system on the rear-end storage system, the study incorporates the grid technology to design a distributed storage method with a load-balancing function. The related gird technology utilizes an improved Ganglia Broker to balance and disperse rear-end computer storage space and system load, and then offers a load-balancing storage framework to resolve the congestion problem as a result of transferring a large amount of data; the system also adopts Globuspsilas RFT components to process data replication and collect all anomaly image files, offering the user with management and surveillance functions.