Hydraulic Oil Pollution Particles Detection Based On Image Processing

Oil pollution is one of the key problems restricting the development of hydraulic technology, the solid particles in the oil are the most common and harmful pollutants in the hydraulic system. The particle detection equipments commonly used at present are either expensive or inconvenient to operate. Based on the analysis of the current technologies for oil pollution detection, this paper makes a deep research for particle detection based on the computer image processing technology. The focus is to identify oil pollution particles in an image and gain related parameters, and then calculate the level of oil pollution. For this purpose, the main works are as follows:1. Study the image preprocessing method for the particle image, adopt cross window filtering method and minimum variance smoothing algorithm to improve image quality.2. Analyze and compare the effect of edge detection algorithm for the particle image, use the improved sobel algorithm to detect the particles edge.3. Propose a segmentation scheme based on the absolute value background subtraction, solve the problem that color difference may cause the loss of information when to be segmented. Adopt image morphology processing method to solve the problem of particle adhesion. Use small area filtering method to fill the region holes. Then gain the connect region characteristics and statistics particle parameters.4. Develop the hydraulic oil pollution particles detection software system in Matlab platform. Through the analysis of the experimental data to verify the effectiveness of the detection method.