How about trying  the dot plot?

e.g.

http://athena.bioc.uvic.ca/virology-ca-tools/jdotter/

Unfortunately, I am not aware of existing software that will do this for you though you might try BEDTools genomecov (see link) which could partially help you.

To my mind this is a sliding window problem, and would be relatively trivial to handle if you have any coding experience. Given a window size of x bp and a step size of y (where y < x, and y is a factor of x), you calculate your statistic (say number of bases within repeats) for a window and move along the sequence y bases to the next window. The data could be easily formatted as say a BEDGraph file (see second link) which you could use in IGV or the UCSC genome browser.

I tend to just keep repeats as features (like a gene feature) until I know exactly what I want to say about something. But i assume that you have something in mind.

Hope this helps. Good luck

http://bedtools.readthedocs.org/en/latest/content/tools/genomecov.html

https://genome.ucsc.edu/goldenPath/help/bedgraph.html

I did not entirely understand what you want to do and I am in doubt among two things: 1. plot the richness in repeated regions

2. characterize the distribution of different repeat families in the chromosome fragment and plot it.

First point can be explored e.g. by identifying repeats with repeat masker or some other algorithm; dividing your DNA segment in windows of a pre-defined size; calculating the fraction of repeated nucleotides with respect to windows size (please see Figure 1 in Surbanovski et al., 2015 on the Plant Journal. There we use the fraction of nucleotides belonging to a transposon, but you can do the same with any repeat.

Second point can be explored using a circos-based plot (http://circos.ca/) or if too complicate to set up, a simple homology based graph to visualize with cytoscape. First record all repeats using your preferred tool. Then extract the corresponding sequences and blast them all-against-all. You can filter the output at whichever stringency you want (e.g. for E-value, identity percentage...). The filtered tabular blast output can be directly interpreted as a graph (in the mathematical sense) that is, every line can be imagined as an edge of a homology graph indicating that two regions share a certain level of sequence identity. Regions that appear more than a certain number of times can be thought as being repetitive. This file can be loaded directly in cytoscape with the network import tool. What you will have in the visualization is nodes corresponds to repeats connected by edges when their similarity is above the threshold.

The same blast output can be adapted with some work to be used as input to circos (or alternative implementations of circular plots). In this case the resulting picture will also contain information on where the repeats are located in the DNA segments, and you will have ribbons connecting regions of the DNA segments on the basis of the similarities detected by blast. see http://circos.ca/intro/genomic_data/  

Circos is a little bit complicated, but very powerful once you dig a little bit into its potentialities.