Ring Feature Extraction In The Application Of Laser Fast Ignition

The research data is obtained from lared_p simulation, which reflect the laser energy variation in the process of transmission in plasma mass.Since the research is established on complex physics background, coupled with the innovative nature of the data itself, it is a quite difficult research domain. After reading lots of literatures, a image feature extract algorithm is proposed about the circle in the image.The simulation data is saved by a time series which can be divided into many time-steps.Each step can be visiualized by image. In the horizontal sections image show many distribution circles, and different circles mean different energy. Under such circumstances circle feature extraction algorithm emerged. This algorithm contains four parts: data pre-processing, edge extracting, edge thinning, circle recognition. Each part contains its own independent technology, But only together constitute the circle feature extraction algorithm. Numerical simulation of a large-scale analysis of the data extracted from the valuable information provided an effective new approach through circle feature extraction algorithm.The pre-processing, as the first part of this algorithm, is responsible for denoising and the foundation of the other parts.Edge extraction stage is used to get the edge information from the slice, which is a very crucial step.In edge thinning stage, an improved skeleton extracting algorithm has been proposed,Through which can solve the long vertical recognition and edge thinning problem. Closed curve in the final stages of identifying an adaptive Hough transform is involved to solve the problem of cirlce recognition.Adaptive Hough transform not only greatly reduced the time and space in the circle recognition, also greatly reduce the influence of noise, which increases the recognition accuracy, in a conclusion, it made a good recognition results.