This paper first describes an efficient image processing based subroutine aimed at assessing railroad ballast wear stage through particle angularity characterization. Three main conclusions were drawn from subsequent experimentations: first, ANG criterion can detect a light wear stage; second, the criterion is very discriminatory; third, ANG values may differ between rock types. Afterwards, the paper focused on the assessment of 13 ballast sources, in order to build a reference angularity database. A comparison with results from the previous experimental program showed an angularity threshold value, below which ballast appears to be worn away. Finally, the angularity of a degraded ballast has been measured and the efficiency of a commonly used recycling process assessed. The angularity of degraded ballast prior to recycling appeared not to be as low as excepted, possibly because of stress localization. An angularity drop was observed after recycling, which may be due to the washing device.