Detection of wells in images of deformed 96-wells plates

Applications that use smartphone cameras for visual colorimetric analysis are often developed in combination with components that serve as sample holders, fluidic chambers, or active membranes. Using smartphones as handheld visual colorimetric analysis devices and sample managers with novel flexible materials raises challenges not present when using rigid materials and better-controlled settings. For instance, accurately finding where analytes are is crucial to a correct analysis. This work has used deformed 96-wells plates as a model of devices that may be used in the field in combination with smartphone applications as colorimetric devices. An algorithm based on the application of the Hough transform followed by an interpolation is developed and tested with plate images that have been deformed in a controlled fashion. The procedure accurately detects the wells in all images of the test sets.