| Summary: | 碩士 === 國立勤益科技大學 === 研發科技與資訊管理研究所 === 102 === Council of Labor Affairs statistic office, most of the reason for the first occurrence of hazardous caused by unsafed behavior, while company set provisions related to the process, when the worker enter the danger zone job site, need to wear protective gear, but still many people seek for their own convenience. Biometric detection technology in the field of face detection applications. Or there is a cherished psychological luck, and forget the good protective measures, therefore other companies have to find another way to reduce the incidence of occupational hazardous.
The proposed system, Field Operator Detection System (FODS), is developed to prevent industrial accident from an unsafety place for field operators with helmet. Based on human face detection technic, the helmet detection function is integrated into FODS detection system which uses a hybrid method to detect field operators with helmet, i.e., Haar-like features, Adaboost algorithm, HHOG algorithm and Support Vector Machine. In the proposed system, the first layer is designed to detect candidate regions form a video frame with Haar-like features and Adaboost algorithm. The second layer of the proposed system, HHOG algorithm and SVM algorithm, is designed to find out optimal candidate regions from the results of the first layer. Therefore, the proposed system can improve the detection rate with two layers structure. The experimental results of the study show that the accuracy rate of field operator detection in video frames can reach 91%. Thus, it proves that our proposed methods can be efficiently applied to field operator face detection.
|
|---|
