Transputer implementation of the Radon transform for image enhancement

The authors briefly describe the Radon transform and the computational considerations necessary to enhance linear features in noisy images. The implementation of the transform a multitransputer network is discussed, and results are presented to demonstrate the performance of the network topology. Feature identification in synthetic aperture radar (SAR) images is considered as an example. The multitransputer network described performs Radon transforms within a few seconds, and linear features in noisy images can be rapidly enhanced. The transputer network architecture can easily be reconfigured to suit various processing algorithms, and the system therefore provides versatile high-speed image processing facilities for a number of computationally intensive applications