It is well known that magnetic resonance imaging (MRI) contrast can be controlled, albeit sometimes at the expense of image resolution and signal-to-noise ratio, and most studies of articular joints have used a single MRI protocol, which is optimised for subjective image analysis. Inevitably that single protocol frequently compromises the detection of one or another of the boundaries between which any measurement must be made. This paper describes an alternative approach in which the criteria for computerised edge detection necessary for fully automated measurement of cartilage thickness are used to define the MRI acquisition parameters. This necessitates the combined use of two MRI sequences, one optimised for the cartilage-bone boundary, and the other for cartilage-synovial fluid. This provides a highly effective combination and its efficacy is demonstrated for the distal interphalangeal joint of a range of asymptomatic adults.