Ultrasonic denoising with a modified wavelet filter

Enhancing signal of interest over a noise level represents a common task for signal processing methods used in measurement systems. Digital filters are used for frequency band limited signals as a common signal of interest enhancement tool. If is the information localised not only in frequency domain but also in a limited time window within the signal, then wavelet processing represents a possible signal denoising solution with an advantage of a higher denoising efficiency over a common digital filter. Development of the wavelets has been focused on signal thresholding methods and adjustment to new sensor systems. We would like to present a fast algorithm for enhancing wavelet denoising performance. This is done by implementing an ideal band-pass filter over common wavelet thresholding process. We propose a new technique for signal denoising which is concentrated on the described ultrasonic signals. The proposed algorithm was verified on ultrasonic sensors which are widely used in the measurement of dimensions and quality of a product. Results showed a significant improvement in wavelet denoising performance over previously published wavelet denoising algorithms.