Object Recognition Using Artificial Neural Network and Discrete Wavelet Transform

Robots and inspection systems use machine vision to increase flexibility and quality in automated production. Due to the increasing expansion of vision systems, various algorithms have been developed to increase the speed and precision of image processing and extraction of its features. In this research, a Wavelet transform based approach is presented for segment classification and identification of its features. therefore, the image of components which are used in the assembly process was captured in different angles and positions. Initially, the 2D DWT was applied to the images. By applying a value of the threshold, the coefficients of the wavelet transform function are selected and using the coefficients obtained from this conversion, the characteristics of the wavelet coefficients are calculated. An artificial neural network of multilayer perceptron was used through the extracted features of the components. To find the best vector characteristics, various combinations of extracted properties from different wavelet stages were investigated. The results showed that the algorithm based on the wavelet transform function has 93.5% accuracy classification.

Keywords:

machine vision , wavelet transform , artificial neural network , object detection feature selection