Abstract:The traditional image recognition method cannot effectively detect the specific position of the power component, and the recognition accuracy is low in the scene with many interferents. In view of the above problems，this paper proposes a method based on MASK LSTM-CNN model for power component inspection image recognition. Combining with the existing Mask R-CNN method,the MASK LSTM-CNN model is constructed by using the long-short-term memory neural network, and the context information is constructed by the fusion of the context information. Then, the optimization algorithm is used to optimize the parameters of the model by combining the specific characteristics of the power components. The model can accurately identify the power components in the field environment with more interference information, and successfully solves the problem that the power components existing in the existing methods have low recognition rate under the occlusion condition, and greatly improves the accuracy of component recognition. Combined with the actual collected power component inspection image dataset, a large number of tests are carried out on the proposed model. The experimental results show that the proposed MASK LSTM-CNN model is better than R-FCN and faster R-CNN. The recognition accuracy rate is improved by 9%-12%, which effectively solves the component identification problem in the power field with more interference information.