Comparative Robustness Evaluation of Code 128 and QR Codes under Controlled Degradation

Abstract

Barcode robustness is determined jointly by symbol configuration, image quality, degradation geometry, and decoder behavior. This study presents a reproducible matched computational benchmark comparing Code 128 and QR Code under Gaussian blur, synthetic low-light noise, glare, perspective and rotational distortion, and simulated physical damage. Twelve deterministic alphanumeric payloads were encoded in both formats and evaluated at five ordered severity levels with three generated replicates per cell, producing 1,800 decoding trials or 900 matched pairs using the same ZBar engine. QR Code achieved a higher overall correct-decoding rate than Code 128 (72.3% vs 68.8%; paired risk difference 3.56 percentage points, 95% CI 0.32 to 6.79; McNemar p=0.037). The difference increased under predefined severe conditions, while performance remained degradation-specific: QR Code was more tolerant of blur, low-light noise, and moderate geometric distortion, whereas Code 128 performed better under the tested glare and distributed-damage models. Two Code 128 misdecodes under extreme distortion illustrate the importance of treating false decoding separately from no-read outcomes. The principal contribution is a severity-dependent benchmark framework rather than a new decoding algorithm or a claim of universal symbology superiority. The findings apply to the tested software-generated configurations and require validation with printed labels, measured lighting, and physical scanners before operational deployment.