As the focus on blockchain scaling research intensifies, the security of light nodes has become a prominent concern. Bitcoin employs Simple Payment Verification (SPV) to ensure the trustworthiness of light nodes by verifying transactions through a proof-of-existence process. Light nodes broadcast confirmation requests to neighboring full nodes, who respond with Merkle proofs for the requested transactions. This system is dependable when the network is predominantly honest, but vulnerabilities arise when an attacker controls a majority of nodes, enabling them to deceive light nodes with fabricated proofs. This paper introduces erasure code to enhance data redundancy and combines it with Merkle proofs to detect protocol rule violations. By integrating coding patterns with random sampling, we verify the integrity of data blocks and enable light nodes to detect fraud by receiving and verifying invalid block proofs from full nodes. We validate the efficacy of our coding-based random sampling model using simulations and probabilistic calculations.