PSI - Issue 84

Regina Finocchiaro et al. / Procedia Structural Integrity 84 (2026) 1264–1269

1266

• Level 3: preliminary or specialized targeted assessments supporting intervention definition and prioritization. • Level 4: detailed assessments and safety verifications, with possible modeling and refinement of the knowledge level. The Level 2, L2, namely the determination of Class of Attention CdA, is identified as the operational fulcrum underpinning subsequent evaluations, because it translates data and observations into actionable priority. According to L1  L2 evaluations, assumed as L2 results, the L3/L4 alternative levels could be defined. The multi-level classification is not purely structural: the overall CdA derives from the logical combination of four components (structural/foundational, seismic, hydraulic, landslide risk levels), each risk level is defined by labeled classes and logical operators. The system is associated to by “operational instructions” clarifying compilation procedures, standard forms and descriptive criteria. Due outlined uncertainty, in November 2025 the CSLP (Italian High Council of Public Works) in cooperation with ANSFISA (National Agency of Road and Rail Infrastructure) approved an update of previous operational instructions, having the declared objective of making application more homogeneous and traceable (without changing the substance of the Guidelines). Within the logic of this present research, two aspects are crucial: a) qualitative and descriptive nature of a significant part of input data (that are prone to operator evaluation especially at L1) b) the crucial role of inspection forms collection (inventory, defectiveness, critical elements) in transition toward L2 CdA definition. In conclusion, the procedure robustness strictly depends on descriptive field-collected data quality: for this reason, Visual inspection is not an instrumental measurement: it is a process of selection, grouping and interpretation of signals, carried out under operational conditions that are often non-ideal (accessibility, lighting, occlusions, environmental context, previous repairs [disregarding operator experience and technical background]). Gestalt provides a useful vocabulary to describe how signs are organized into meaningful units. In technical practice, this coincides with the transition from “defects” to “phenomena” like cracking patterns, diffuse deterioration, localized effects and possible cause–effect interactions. The literature on inspection reliability highlights variability across observers and across inspection campaign approaches, with dependence on experience, context and cognitive workload. In this framework, to make explicit certain aspects of perceptual grouping does not mean a psychologizing engineering approach; it means improving reproducibility, traceability and quality of CdA information and, especially, robustness of any further investigation decisions. 4. Perception effects on multi-level procedure Even when standardized forms and defect taxonomies are available, there are recurrent operational points at which perceptual organization inevitably shapes the process outcome and consequently the final result. A first critical node concerns the passage from recorded defectiveness to identification of a “phenomenon”. The inspection forms typically document point-like or local signs, yet the operator must decide whether multiple observations belong to a single coherent process or to distinct, unrelated, processes. This decision is intrinsically perceptual and highly sensitive and experience based, since grouping tends to be driven by proximity, similarity and perceived continuity, which can lead to a classical bias: either over-aggregation (merging independent defects) or excessive fragmentation (splitting a coherent pattern). A second node involves the identification of critical elements. Human attention is naturally attracted to what visually stands out, either because of contrast, singularity, or locally striking evidence. As a consequence, low-salience but spatially diffuse deterioration may be systematically underweighted, despite its potential relevance to vulnerability, durability, and long-term performance. This introduces a structural bias in the allocation of attention and, therefore, in subsequent prioritization of observations. it becomes necessary to introduce “ perceptual quality ” in a stable manner. 3. Perception and organization of detected defects: the role of human factor