The application of Hopfield neural network in enhancing x ray image of steel pipe welding

This paper analyses the characters of x-ray image of thick and thin steel pipe. In order to enhance the x-ray image automatically and avoid deciding the image´s degraded type, a gray mapping matrix is constructed to replace traditional gray transformation curves and the maximum dimension of the gray mapping matrix is 256×256. So the calculation time has little relation with the size of the image. The criterion function of image quality is used to evaluate the quality of the transformed image. By this way, the problem of image enhancement is transformed to an optimization problem. The paper presents Hopfield neural network to calculate the gray mapping matrix. The energy function and the calculation method are also given. Some examples show that the presented method is effective.