Implementation and Hardware Acceleration of Wiener filter algorithm using Vivado High Level Synthesis HLS for X-ray baggage Images

Abstract

Many computer vision applications require a real-time image de-noising technique. In recent years to detecting hazardous materials using X-ray baggage images is important means of airport security applications. The image denoising functions to aid decision making by humans is a predictable area of critical need. However, implementing the image processing approach in Hardware Description Language is time consuming and difficult to debug. With High Level Synthesis HLS, the designer may take advantage of libraries to reap the benefits of working at a higher level of abstraction while producing high-performance hardware. HLS tools, on the other hand, provide prospective solutions to hardware design. HLS tools provide us with extensive design assessment capabilities as well as a wide range of optimization strategies. In this research article, Intellectual Property IP core of wienerfilter algorithm is built utilizing HLS and implemented on ZYNQ platform for X-ray baggage image screening applications. We used Pipeline Pragma techniques added to the HLS environment improved performance, and the results were far better than predicted. The solution using the Pragma method with pipelining was to optimize area and 89 percent clock cycle reduction, and the relevance of the pipeline with this speed was well suitable to improve real time performance.