Research Of Sequence Control Method For The Equipment With Image Recognition

With the rapid development of Internet and artificial intelligence,its related technologies are more and more widely used in smart power generation,smart power plants and other fields.Through the intelligent software,the centralized control operators automatically carry out the corresponding equipment operation in accordance with the pre-established equipment operation rules,which is a new idea of smart power plant research and power plant upgrading.The work contents and problems encountered by the centralized control operators in the power plant are analyzed.In the process of executing the equipment operation tasks sequentially in DCS,the specific equipment operation and interface observation behaviors of the operators are abstracted into a sequence control model based on image recognition.Based on this model,an equipment sequence control method is proposed,and the equipment sequence control system based on image recognition is realized,which includes voice control,equipment sequence control logic judgment,log recording and other functions.the system can replace the operator to complete part of the operation task.In this paper,the operation instructions and operation tasks commonly used in the process of centralized control operation are summarized and sorted out,and the voice instruction library of the system and the equipment sequence logic of operation tasks are designed.The speech control function consists of two parts: speech recognition and humanoid operation.Based on Microsoft Speech API 5.1,the speech instruction recognition is realized,and the humanoid operation of keyboard and mouse during the execution of the device is reproduced.The function of equipment sequence control logic judgment is to identify the interface information such as the current DCS operation interface and the status of the equipment to be operated,instead of the operator to judge the sequence control logic conditions of the operation task.Based on MSER region edge enhanced title extraction and text recognition based on CRNN network,the title recognition of DCS operation interface is realized.Based on the device extraction method of background elimination,the image of the operated device and its position are obtained.Through the image similarity method of fusing local and global features,combined with the location judgment of the device,the accurate recognition of the device is realized.By matching the color recognition with the digital template,the identification of the status of the device and related parameters is realized.In addition,the logging function is written to realize the device operation logging function.Finally,by simulating the operator to perform the operation task in the DCS interface,the function of the equipment sequence control system is tested.the test results show that the system can receive instructions,carry out the corresponding equipment sequence operation and related logic judgment,and record the operation log to complete the operation task.