Nonlinear spatial transforms and fuzzy pattern classification with unimodal potential functions are established in signal processing. They have proved to be excellent tools in feature extraction and classification. In this paper, we will present a hardware-accelerated image processing and classification system which is implemented on one field-programmable gate array (FPGA). Nonlinear discrete circular transforms generate a feature vector. The features are analyzed by a fuzzy classifier. This principle can be used for feature extraction, pattern recognition, and classification tasks. Implementation in radix-2 structures is possible, allowing fast calculations with a computational complexity of O(N) up to O(N·ld(N)). Furthermore, the pattern separability properties of these transforms are better than those achieved with the well-known method based on the power spectrum of the Fourier Transform, or on several other transforms. Using different signal flow structures, the transforms can be adapted to different image and signal processing applications

No abstract

No abstract

No abstract

No abstract

No abstract

No abstract

Lifetime and reliability of high-power diode laser bars are sensitively related to operating temperature, mounting stress, and solder electromigration. These three factors have been taken into account for the development of a new packaging technology for 1-cm laser bars of gallium arsenide. We examinae the use of chemical-vapor-deposited (CVD) diamond as heatspreaders in order to reduce thermal resistance of a microchannel cooler for liquid cooling. We show that it is possible to perform hard soldering on a CVD-diamond with a new technique. Additionally, we present a controlled water cooling system fit to the flow characteristics of the cooler. It permits one to adjust the emission wavelength of the diode lasers by changing the water flux