Vision Processing with the Canny Edge Detection Reference Design

ID 683433
Date 2/14/2015
Document Table of Contents

1.4.1. Greyscale Conversion

Edge detection algorithms operate on one color channel. The video input has three color channels, so the design converts it to grayscale.

The design uses the NTSC standard color space to convert. The design avoids using complex floating-point hardware, which limits fMAX performance and introduces clock cycle latency if pipelining is used to increase fmax performance. The design uses bit-slice method to simplify floating point calculations where precision is not required. However, this method can only be used if the coefficients are constant.

Figure 2. Greyscale