In digital forensics, proper evidence management is crucial for police investigations due to its volatility and improper handling technique. However, to present and utilize the captured digital forensic images, which are typically concrete and admissible in court and a trial, is very challenging. In addition, centralized evidence-gathering, classification, and archiving techniques diminish the credibility of digital evidence. To address these issues, this work proposed a blockchain-based classified digital forensic sensitive image management framework. The framework includes a rapidly expanding polygon chain, Interplanetary File System (IPFS), and a deep learning model that performs with more than 98% accuracy. This framework preserves sensitive forensic images in the fully decentralized blockchain. Our system uses metamask wallet-based authentication to guard against unauthorized users. We use IPFS to produce the image hash to encrypt and save the forensic image in the blockchain. A voting method based on smart contracts is suggested for accessing the preserved image. We implemented the framework, which offers promising performances in response time, evidence insertion time, evidence access time, communication overhead, and effective forensic image categorization.