In this paper, a piezoresistive pressure sensor based on silicon on insulator (SOI) was presented, which was composed of an SOI layer with sensing elements and a glass cap for a hermetic package. Different from its conventional counterparts, the position and thickness of the four piezoresistors was optimized based on numerical simulation, which suggests that two piezoresistors at the center while the other two at the edge of the pressure-sensitive diaphragm and a thickness of 2 μm can produce the maximum sensitivity and the minimum nonlinearity. Due to the use of silicon rather than metal for electrical connections, the piezoresistive pressure sensor was fabricated in a highly simplified process. From the experimental results, the fabricated piezoresistive pressure sensor demonstrated a high sensitivity of 37.79 mV·V⁻¹·MPa⁻¹, a high full-scale (FS) output of 472.33 mV, a low hysteresis of 0.09% FS, a good repeatability of 0.03% FS and a good accuracy of 0.06% FS at 20 °C. A temperature coefficient of sensitivity of 0.44 mV·MPa⁻¹·°C⁻¹ and a low zero drift were also shown at different temperatures. The piezoresistive pressure sensor developed in this study may function as an enabling tool in pressure measurements.

The constraints of traditional wavelet filters and wavelet coefficients such as it can not be based on the input signal changes, combined with underwater acoustic channel transmission and imaged sonar principles causes the sharp edges or the presence of singular points of piecewise smooth sound effects far from ideal. To deal with these shortcomings, this paper proposes an improved SPIHT algorithm which image most of the energy is concentrated in the low frequency sub-band after wavelet transform. Though SPIHT algorithm diversity redundancy judgment rule which is located in the low frequency sub-band Wavelet coefficients resulting in a code scanning initial zerotree and producing more unnecessary sign bit is relatively large. Using the methods and diversity judgment rules of changing the original algorithm in wavelet tree construction, which can reduce the initial stage of the zero-tree scan number and redundancy of diversity judgments resulting as well as improve the coding efficiency. Matlab language is used to achieve the analysis of underwater acoustic image compression. As the comparing results showed that the improved adaptive wavelet transform algorithm reduce the wavelet transform coefficients to obtain more ideal compression ratio at a certain extent. To ensure the stability of the transformation process, the acoustic image compression method has a good application value.