Improving the Accuracy of Image Segmentation and Noise Reduction Based on Artificial Intelligence and Wavelet Transformation

Main Article Content

Hu Pingfang

Abstract

Laser imaging is interferedwith by the environment, equipment, etc, making the laser images contain noise. The
current image segmentation methodshave poor robustness for noise interference, a high probability of false
segmentation, and serious loss of important information. Therefore, in order to overcome these disadvantages, a
laser image segmentation method based on the deep learning of artificial intelligence was put forward.First of all,
the wavelet transform was used to carry out feature extraction of the laser image, and then suppress the noise
interference. Later, the artificial intelligence learning algorithm was introduced to train the feature vector of the
laser image, and the laser image pixels were classified according to the training results, so as to realize laser
image segmentation. Finally,the laser images with and without noise were used forthe simulation test.The results
indicated that the segmentation accuracy of artificial intelligence deep learning for laser images with and without
noise was 91% and 95%, respectively, which was significantly higher than that of the classical laser image
segmentation method, so the segmentation efficiency could meet the requirements of large-scale laser image
development.

Article Details

How to Cite
Hu Pingfang. (2021). Improving the Accuracy of Image Segmentation and Noise Reduction Based on Artificial Intelligence and Wavelet Transformation. CONVERTER, 2021(7), 988-994. Retrieved from http://converter-magazine.info/index.php/converter/article/view/586
Section
Articles