Grain storage security is a crucial component of food security, and adopting advanced technical methods is essential for effective supervision. In recent years, radar technology has become a prominent focus in grain storage monitoring. This study utilizes microwave transmission technology to measure the complex dielectric constant of grain and establishes a calculation model linking grain density and dielectric constant, offering a feasible approach for detecting the average density of grain piles. To address the “underdetermination” issue in identifying anomalies using ground-penetrating radar (GPR) echoes, a finite-difference time-domain (FDTD) simulation model of the granary detection environment is developed. By integrating simulation results with actual 2D radar echo images, high-precision positioning of anomalies within the granary is achieved. Electromagnetic detection experiments further validate the feasibility and practicality of cross-hole radar technology for identifying and imaging foreign objects in grain bins, resulting in a practical mathematical model. The model achieves a positioning accuracy of 0.3 m, meeting engineering application requirements. This research offers a scientific and reliable method for non-destructive detection of foreign matter in grain bins, providing new insights and practical solutions for improving grain storage monitoring.