Abstract: In order to solve the problem that the fluctuation of feed particle size affects the processing efficiency of subsequent processes in the process of ore grinding，a system of ore block identification and autonomous switching of feed path is proposed. The hardware composition，communication network，image processing flow，control strategy and implementation method of the system are described.Through the use of six parallel ore-feeding belts and one main belt，the particle size is divided into fine-grained (No.1， No.2)，moderate-grained (No.3，No.4) and coarse-grained (No.5，No.6). The six belts and the main belt are equipped with a block analyzer，combined with the watershed algorithm for image processing and parti⁃ cle size calculation，real-time detection of F80 indicators and block analysis is carried on.The electronic belt scale monitors the processing capacity through the Modbus RTU protocol，and the block analyzer com⁃ municates with the DCS through the OPC protocol. The control system automatically selects the appropriate grade of the ore-feeding belt according to the target processing capacity and F80 data，and adjusts the run⁃ ning speed of the feeder through the PID algorithm to achieve uniform and stable feeding，which can pro⁃ vide an effective solution for the intelligent control of the ore crushing process.

Key words: block identification, block analysis, watershed algorithm, PID control, Modbus RTU, OPC