Vehicle door sub-health diagnosis method and system based on image processing

The invention discloses a vehicle door sub-health diagnosis method and system based on image processing. The method comprises the steps that vehicle door operation signals in a health state and different types of sub-health states are collected and subjected to image processing; adopting a Dense-sift algorithm to carry out feature extraction on all the sub-health grey-scale maps after imaging processing, and determining a plurality of clustering centers; then Dense-sift feature vector distribution is carried out based on the clustering center, and word bag features of different types of sub-health states are obtained; an improved outlier detection method is adopted, and interference word bag features are eliminated; and constructing a sub-health diagnosis model based on the word bag features of each type of sub-health state after feature screening and the sub-health state type labels, and performing diagnosis. According to the method, the sub-health diagnosis model constructed after invalid features are remov