High level estimation of implementation cost of using morphological filter in edge detection

Abstract

Edge detection is a critical element in image processing, since edges contain a major function of image information. The function of edge detection is to identify the boundaries of homogeneous regions in an image based on properties such as intensity and texture. The morphological filter is used as an initial process in edge detection for noisy images where “opening-closing” operations are used to filter noise thus enhancing edge detection performance. In this paper we study the additional cost in resources caused by implementing the morphological filter prior to edge detection, area, power and energy consumption are considered, the cost is a major factor in determining to use or not the morphological filter in a particular application. To achieve this estimation we used a high level estimation tool, high level design and estimation is gaining ground as it allows design decisions in an early stages of development therefore reducing costs, these tools are also gaining in accuracy. We used morphological filter as a preprocessing stage for the Shen-Castan edge detector. Starting C code the high level estimation tool produces RTL level circuits and using power and energy consumption models based on a hardware database (generic ASIC in our case) it produces reports about Area, Power and Energy consumption, an estimation of performance is also possible.