Fault diagnosis for snapshot device through identifying image abnormal features

Traditionally, faulted snapshot device is diagnosed artificially, which is inefficient and has low accuracy. This paper proposes a method of fault diagnosis for snapshot device through identifying abnormal images caught by the device. First, nine types of image abnormal features are classified and described, abnormal features are also mapped to different faults of snapshot device. Second, the distribution of three channels, R, G and B of original color image caught by the snapshot device are calculated and analyzed , the features of mean gray value,distribution of gray value and region features in gray image are analyzed, white light beam in binary image,as well as aspect ratio are both researched deeply. Third, different algorithms are studied to identify different image abnormal features and then the whole diagnosis method is developed. Finally, the proposed method is proved to have a high efficiency and accuracy in a case.