Key Features Extraction From Algae Proliferation Process And Its Application

Traditional eutrofication monitoring systems can not predict or previously warn harmful algal bloom (HAB) because they were based on measuring the color and area of destroyed water body, thus can only mornitoring HAB after it has occoured. Therefore, exploiting a simple, high-accuracy, low-cost prediction and warning system is a major direction of eutrophication research. In order to provide basic references for previous warning technologies, in this thesis, harmful algae identification and quantitative prediction were stutied through traditional means of microscopic observation plus computer image analysis.The key morphological and photosynthetic characteristics were extracted through the study on the variation of morphology, number, photosynthetic oxygen production rate during proliferation of algal taxa Microcystis aeruginosa. The result was verified by comparing with algae microscopic images. Correspongdinly, neural network technology was utilized to stimulate the algal proliferation. In addition, in order to verify the feasibility of this method, the proliferation of algae as well as the changes of algae spicies and physiological ecology characteristics before and after the water bloom were monitored in artificial enclosure with M. aeruginosa bloom. The main conclusions are shown as follows:(1) There are four phases during M. aeruginosa proliferation: lag period (1-3d), logarithmic growth period (4-20d), stabilization period (21-25d), aging period (26-29d). Specific growth rateμof the four proliferation phases were increasing rapidly in 0.4-1.0d^-1, stability changes in 0.2-1.0d^-1, a declining trend in 0.0-0.1d^-1, little change in almost <0d^-1, respectively. Division rates of the four proliferation phases were increasing rapidly in 30%-50%, changes unsteadily in 30%-50%, stability changes in 10%-15%, increased trend slightly in 15%-30%, respectively. Cells areas were increasing rapidly in 35-120 pixels, stability changes in 80-160pixels, increasing slightly in 60-120pixels, increased trend in almost >120pixels, respectively. Chla concentration of M. aeruginosa at logarithmic growth period, stabilization period and aging period were 7.89×10-8, 9.06×10-8 and 4.90×10-8μg/cell, respectively.(2) Different respiration rates and light compensation point in different proliferation phases. M. aeruginosa at logarithmic growth period has stronger low-light adaptability than others. At the same period, the higher temperature, respiration rates are stronger. At the same temperature, respiration rates are highest at aging period. The respiration rates at logarithmic growth period, stabilization period and aging period were 45,37,55μmolO₂·mg^-1·Chl a·h^-1 respectively under 20℃and 67,57,117μmolO₂·mg^-1·Chl a·h^-1 respectively under 25℃.The light compensation point were 250,900,780lx respectively under 20℃and 380,720,480lx respectively under 25℃.(3) Endogenous metabolic factor K of Scenedesmus obliquus are different under different temperature: temperature increases, K-value increases and light intensity increase, K-value decreases. The K-value at 0,200,500lx is 0.20,0.17,0.11d^-1 respectively under 20℃and 0.24,0.21, 0.23d^-1 respectively under 25℃. Different K-value in different algae species. The K-value of Microcystis aeruginosa, Scenedesmus obliquus and Auabaena Flosaquae were respectively 0.37,0.20,0.30d^-1.(4) After Microcystis aeruginosa bloom was induced in artificial enclosure, the light shading material was covered on the surface and the variation of chlorophyll a contents, pH and DO were analyzed. Chlorophyll a, DO and pH declined from 107.1μg/L, 9.7mg/L and 9.1 to 44.5μg/L, 2.6 mg/L and 8.0 respectively. Meanwhile, microcystis colony's radius decreased obviously, from 70% larger than 50μm to less than 50μm. It was shown that microcystis colonies floated upward to the water surface in light shading by means of monitoring Chlorophyll a contents at different water depths,. The buoyancy control mechanism of Microcystis aeruginosa may explain this phenomenon.(5) Succeed in predicting the growth trend of Microcystis aeruginosa proliferation by simulating morphological characteristics and growth curve using kinds of neural network models.