blobify: Denoise a black and white image to only core features

In niconaut/crop.row.finder: Crop Row Detector

Description

This takes a black and white image, preferrably the image produced by an otsu transformation, and gets rid of all extrenuous features. Many long skinny features and random clumps of features will be erased, leaving only the larger core features of an image.

Usage

blobify(image, size)

Arguments

image	An image.
size	An odd number, the starting size of the kernel for morphology, minimum of 3 starting size.

Details

Using morphology functions https://en.wikipedia.org/wiki/Mathematical_morphology to get rid of unwanted noise and features. With this package the features to eliminate will mostly be random weeds and bushy leaves. The specific functions being use are opening https://en.wikipedia.org/wiki/Opening_(morphology) and closing https://en.wikipedia.org/wiki/Closing_(morphology). The first step of the function is opening to start to get rid of smaller features. It then closes the holes created by opening the image and repeats this process several times with bigger and bigger kernels. This allows more important featrures to remain while insignificant ones are erased. There is a minimum starting size of 2, because when doing the first closing process it needs to be at least a 1 pixel kernel.

Examples

final_blob <- blobify(image, size)

niconaut/crop.row.finder documentation built on June 29, 2020, 6:10 a.m.