Design of modified Canny Edge Detector based on FPGA modified Canny edge detection algorithm

A modified Canny edge detection algorithm
that adaptively computes the edge detection thresholds
based on the block type and the local distribution of the
gradients in the image block. The design
of modified Canny Edge detection algorithm that results
in significantly reduced memory requirement, decrease
in latency, increase throughput, with no loss in edge
detection performance as compare to original Canny
Detector Algorithm.

In order to reduce memory requirements, decrease
latency and increase throughput, a distributed Canny edge
detection algorithm is propose. The hysteresis threshold
calculation is a key element that greatly affects the edge
detection output. However, the original Canny algorithm
computes the high and low thresholds for edge detection
based on the entire image statistics, which prevents the
processing of individual blocks independently. A new threshold selection algorithm based on the distribution of pixel gradients in a block of pixels to overcome the dependency between the blocks. However, the hysteresis
thresholds calculation is based on a very finely and uniformly
quantised 64-bin gradient magnitude histogram, which is
computationally expensive and, thereby, hinders the real-time
implementation. In this paper we are proposing compression
of image and then applying distributed canny edge detector
algorithm in VHDL.

Smooth the image with a 2 dimensional Gaussian. In most
cases the computation of a 2 D Gaussian is expensive,
therefore it’s approximated by 2 one dimensional Gaussians,
one within the x direction and also the alternative within the y
direction.

Take the gradient of the image. This shows changes in
intensity, that indicates the presence of edges. This truly
provides 2 results, the gradient within the x direction and also
the gradient within the y direction. Non-maximal suppression. Edges can occur at points the wherever the gradient is at a most. Therefore, all points not at a most should be suppressed. So as to try and do this, the
magnitude and direction of the gradient is computed at every
picture element. Then for every picture element check if the
magnitude of the gradient is larger at one pixel’s distance
away in either the positive or the negative direction
perpendicular to the gradient.
The strategy of thresholding utilised by
the cagey Edge Detector is referred to as “hysteresis”. It
makes use of each a high threshold and an occasional
threshold. If a picture element includes a price on top of the
high threshold, it’s set as a position picture element. If a picture
element includes a price on top of the low threshold and is
that the neighbour of a position picture element, it’s set as a
position picture element further. If a picture element includes
a price on top of the low threshold however isn’t the
neighbour of a position picture element, it’s not set as a
position picture element. If a picture element includes a price
below the low threshold, it’s ne’er set as a position picture
element.