Cmc chip-driven anchor machine remote controller fault image diagnosis system

Abstract

The present application relates to computer vision and industrial equipment fault diagnosis technical field, specifically to CMC chip drive's excavator anchor machine remote control fault image diagnosis system;Including data perception module, synchronous acquisition CMC chip internal logic state vector and real-time observation image, and preprocessing obtains target vision data;Generation formula reconstruction module, the logic state vector is mapped to high-dimensional space, and the ideal feature map is generated;Visual analysis module, through convolutional neural network, the observation feature map is extracted, and the interference distribution mask is generated based on the frequency domain texture statistics, and the unstructured environment noise area is identified;Difference analysis module, ideal feature map, observation feature graph and interference distribution mask are combined to calculate semantic residual graph, and image difference analysis result is generated;The present application generates ideal characteristics by using CMC internal logic state vector, and the residual calculation of coal dust noise area is shielded through interference mask, effectively eliminates the false alarm caused by environmental noise, realizes the closed loop verification of seen.

Images