Design and test of a robustness evaluation system for micro-vision tracking algorithms

Industrial applications of micro-vision tracking algorithms become increasingly prevalent. Unfortunately, out-of-focused-plane (OFP) disturbances negatively impact the in-focused-plane (IFP) tracking accuracy of the micro-vision. This paper proposes a robustness evaluation system for micro-vision tracking algorithms. The relationship between IFP accuracy degradation/improvement and OFP disturbances is quantified. First, a commercial spatial nanopositioning stage (com-SNPS) and an SNPS designed in the laboratory (lab-SNPS) were employed to build a robustness evaluation system. Two SNPSs were utilized to generate both IFP trajectories and specific OFP disturbances. Capacitive sensors were used to evaluate the IFP accuracy of micro-vision tracking algorithms. Second, traditional micro-vision tracking algorithms were selected. The combination of the constant-template matching method, constant-region-of-interest (constant-ROI) retrieval method, and constant-focused-plane focusing method acted as test examples. Third, robust micro-vision tracking algorithms were developed. The variable-template matching method, variable-ROI retrieval method, and variable-focused-plane focusing method were combined. Finally, the prototype of the proposed robustness evaluation system was tested. The focused plane was determined to be a benchmark for calculating OFP disturbances. The IFP accuracy of chosen algorithms under specific OFP excitation was measured. Test results demonstrate different IFP degradation or improvement characteristics of micro-vision algorithms. This paper contributes to developing a robust micro-vision tracking algorithm.