| Most of the production lines of bromide in China were built in 1980 s,but there are some problems,such as unreasonable industrial process,backward control system and low extraction rate of bromide.In the actual detection of bromine content,technicians judge whether the titration has reached the end point through color mutation in laboratory at first,and then adjust the parameters on the production site according to the corresponding calculation results.However,the time efficiency of manual assay is poor,and the time interval is long.Therefore,the production process has the serious adjustment lag,and has caused a large amount of waste of raw materials.In addition,there are large random errors in manual detection and naked eye reading,which will seriously affect the accuracy of bromine content detection results.According to the above problems,a set of bromine detection system based on computer vision was proposed to optimize the current production line for bromine detection by analyzing the demand for detection of bromine content in oxidized brine.The system is divided into three parts: image processing algorithm,visual detection system hardware and corresponding software.The hardware of detection system is divided into image acquisition system and liquid conveying system.Among them,the hardware of the image acquisition system is responsible for the image acquisition of the reaction area and ensuring that the acquisition process is not interfered by the external light.The hardware of the liquid conveying system consists of two parts,including the controller with STM32F103 chip as the core and the automatic pipetting device.From the hardware level,the functions of automatic control of visual system conveying,non-contact automatic pipetting and automatic start-stop of intermediate relay are realized.The visual system software includes four parts: automatic analysis and control system of oxidized brine,upper computer of automatic analysis equipment,slave computer logic control and communication program design.The upper computer program is responsible for transmitting the instruction of image processing in the reaction area to the slave computer through the RS485 electrical interface.At the same time,part of the data is also uploaded to the automatic analysis and control system through the RS485 electrical interface after the end of a detection process.The program design of the automatic analysis and control system realizes the information configuration and saves the data transferred from the upper computer to the project folder.At the same time,in order to visually monitor the changes of the data,the data monitoring interface is presented in the form of line graph.The slave logic control program mainly realizes the logic control of STM32 controller and the logic control of injection pump.Communication program designs from the aspect of software to realize the automatic analysis between upper machine and control system,and the information exchange between PC and from the machine,and image processing algorithms are analyzed theoretically.Meanwhile,the algorithm will be extracted from the target area of reaction zone,according to the characters of the color of the computer vision technology and oxidation chemical reaction characteristics of brine,converting target area from RGB to HSV color space,and then judging whether the threshold value of the target area is tonal H in certain range,the threshold directly determines from the machine will execute the next steps.In addition,this paper also tested the function of the bromine content detection system to verify its feasibility.During the testing process,each functional module of the system met the requirements of this topic.At the same time,the overall experimental platform of the visual detection system was built.By comparing the volume of sodium thiosulfate solution titrated by the system and the manual in two different ways,the relative error between the two solutions was obtained within a reasonable error range. |
PDF Full Text Request