Research And Development Of Domestic Sewage Source Detection System Based On Image Recognition

Domestic sewage source separation and resource utilization technology has become one of the research focuses in the field of sewage treatment in recent years.Although the existing finished equipment can detect the sewage quality,the intelligence level is still not high,and the reliability and effectiveness of the detection data is difficult to be guaranteed.With the development of computer video image processing and analysis technology,the application of the technology in production and life has become more and more mature.Therefore,the image sewage detection system can overcome the shortcomings of traditional detection methods.Based on the image recognition technology,the core algorithm of sewage intelligent detection is designed,and the hardware and software system of the intelligent detection system is developed by using Raspberry Pi development board.The main contents and conclusions are as follows:(1)Aiming at the problem of low recognition rate of sewage with single feature and single classifier,a domestic sewage recognition algorithm based on support vector machine(SVM)and D-S(Dempster Shafer)evidence theory is proposed.First of all,we need to extract the static color features of water quality image,train SVM classifier and use it to classify the water quality image to be tested,then use the frame difference method to detect the number of impurities in the image,and then use the user-defined function to convert the number of impurities in the image to its probability of belonging to clean water and polluted water,as the classification result of the dynamic features of the image.Finally,the D-S evidence theory is used to fuse the static features of the image to be tested from SVM and the classification results of the dynamic features from the frame difference method,and finally judge the water quality image.The experimental results show that the average recognition time of the algorithm is about0.08 s,and the average recognition rate is more than 80%.The algorithm not only realizes the real-time detection of sewage,but also significantly improves the detection accuracy.(2)A set of domestic sewage source detection system based on image recognition is developed.The main function modules of the system include image acquisition module,image recognition module,image storage module and image display module.Its hardware development environment is composed of 1.4GHz 64 bit quad core cortex-a53 processor,raspberry PI 3B + development board,ov5647 model camera and TF memory card;its software development environment is composed of raspbian operating system,vnc remote login software,opencv library and Qt development software.Finally,the system is tested,the results show that the whole system can process 25-30 frames per second,and the recognition accuracy is more than 90%.