The invention relates to the related technical field of image recognition, in particular to an internal mixer rotor defect detection method and
system based on image recognition and a medium, and the method comprises the steps: extracting
frequency domain and
time domain features of an internal mixer rotor according to vibration signals, optimizing
image matching defect visual vectors, configuring vibration
signal feature vectors, and evaluating defects through a defect classification network after fusion. And if the result exceeds the limit, triggering reminding and linking the mechanical arm to stop. The technical problems that when an internal mixer works, due to oil
contamination coverage, imaging on the surface of a rotor is fuzzy, the rotor is easily interfered by
mechanical noise in a high-temperature and high-
noise environment, and the defect detection precision of the internal mixer rotor is limited are solved,
frequency domain and
time domain features of vibration signals and optimized rotor image visual features are fused, a defect classification network of an attention mechanism is introduced, and the defect detection precision of the internal mixer rotor is improved. The method has the technical effects of dynamically focusing a tiny defect area, eliminating oil
stain interference, adapting to contrast optimization to cope with a high-temperature environment, still keeping clear imaging in an oil
stain environment, and ensuring the defect detection precision of the internal mixer rotor.