Feature extraction and image recognition for the electrical symbols based on Zernike moment
Yongfei Yin, Zhaoyong Meng, Shaoqian Li · 2017
In nowadays, the image recognition is widely applying to various fields. With the construction of informatization and intellectualization highly developing, it will inevitably becomes investigative trend that apply image recognition to drawings and literal data of power system. This paper mainly studies the Zernike moment that applied to feature extraction of electrical engineering drawings, which will extract features of the electrical equipment symbols in it and recognize. Aiming at the resample and re-quantization matter caused by image rotation and scaling transformation, the improved Zernike moment method was proposed, that is, first target area in the image was normalized shapely, and then the Zernike moments were normalized. After image segmentation of electrical engineering drawings which has been preprocessed, extracting the effective image features. In another word, it means extract the Zernike moment feature set and the improved Zernike moment feature set of different orders to reflect the shape feature of electrical equipment in electrical engineering drawings. Comparing the recognition of the Zernike moment feature set with the improved shows that the improved Zernike moment has an outstanding ability of anti-noise and higher recognition rate.