| Based on the background of practical library condition, the system is mainly based on software system design.The paper focuses on the design of library-IMS to make up for the unsteady and low efficiency of man monitoring system. The detailed contents of the paper mainly conclude implementation of security monitoring module, video capturing module, database management module, communication module and user interface module.1. Abnormal target detection based on improved background differential framework. The specific abnormal detection and skeleton extraction is fulfilled firstly, and then the idea of background refresh is implemented based on the background image quality evaluation system within the improved framework, which also covers the local adaptive threshold-based binarization procedure. The final detection accuracy and extraction quality is improved to a certain extent.2. The abnormal recognition system is achieved based on the HOG(Histogram of Oriented Gradient) feature extraction and SVM(Support Vector Machine) algorithms. In the design of the improved framework, the subject’s shape-based feature representation idea is proposed to enable the detection system with high classification results.3. Demographic based on the local feature extraction and recognition system. It realizes the subject recognition and number record in the reading room. As for the implementation of the subject recognition under the complicated background, the background refresh of the ROI and preliminary sifting is realized to ensure the final counting accuracy.4. The design of the software system. The database management module and communication module is designed into the integrated system, which is capable of storing and transmitting information simultaneously, followed by obtaining the simulated results of algorithms. The practicability of the designed system is well verified by the experimental results and corresponding discussion. |