This research endeavors to establish a methodology for forecasting the sizes of anomalies exceeding normative thresholds in elongated structures such as roads, railway tracks, pipelines, and bridges. While substantial advancements have been achieved in prognosticating the maintenance needs of elongated objects, the effective application of time series prediction techniques for estimating the dimensions of defects remains constrained. The proposed approach entails a statistical examination of a multidimensional feature space, formulation of input, output, and control parameters, construction of a two-dimensional grid, and projection of defect dimensions within the grid framework. The final phase integrates a linear regression model with Ll regularization, demonstrating practicality in predicting defect sizes while mitigating overfitting and enhancing model generalization through complexity penalization. The resultant model offers early detection capabilities for potential issues and facilitates timely preemptive maintenance actions by forecasting defect dimensions based on an extensive array of features.