本文選題：微藻 + 生物膜量��； 參考：《重慶理工大學》2017年碩士論文

【摘要】：微藻的高效培養(yǎng)是生物能源開發(fā)利用的前提及關鍵,目前關于微藻培養(yǎng)方法有懸浮培養(yǎng)和固定化培養(yǎng)。雖然微藻懸浮態(tài)培養(yǎng)模式的操作方便,但采收困難、復雜而且采收的成本比較高;而微藻固定化培養(yǎng)模式的操作穩(wěn)定、生物量密度高,同時采收便利快捷、能耗低等優(yōu)勢,目前已成為微藻領域研究者關注的焦點。雖然微藻固定化培養(yǎng)具有諸多優(yōu)點,但是微藻品質,尤其是微藻生物量受生物膜生長因素影響顯著。因此,開發(fā)出一種快速簡便的無損在線檢測方法對微藻生物膜生長信息進行監(jiān)測,以便于及時調(diào)節(jié)生長條件來提高微藻生物膜量尤為重要。高光譜成像技術作為一種新型的無損、快速、準確的檢測技術,該技術結合了計算機視覺和光譜檢測兩種技術的優(yōu)點,能夠很好的記錄生物膜生長的豐富信息,可為微藻生物量的快速無損檢測找到了一條簡便有效的方法。本文利用高光譜技術與生物測量技術結合,對不同生長環(huán)境條件下的微藻生物膜反射光譜特征與生物膜量預測兩方面分別進行研究分析,提出了一種基于高光譜的微藻生物膜量的監(jiān)測方法,為微藻生物膜生長信息的快速獲取提供了技術支持以及微藻生物膜的高效培養(yǎng)提供了參考。取得的研究結果如下:(1)培養(yǎng)了不同溫度、不同PH值以及不同光強條件下的微藻生物膜。(2)根據(jù)高光譜成像儀的理論基礎,搭建了高光譜圖像采集系統(tǒng),并對微藻生物膜生長信息進行檢測,獲取了不同溫度、PH值以及光強條件下微藻生物膜的高光譜信息。(3)根據(jù)微藻生物膜的原始光譜反射值有三個明顯的特征峰,特征波長(789nm、811nm、930nm)處對應的光譜反射值隨培養(yǎng)時間的增加而降低。(4)根據(jù)采集到微藻生物膜的光譜反射數(shù)據(jù),通過衡量生物量的光譜特征變量,提取出了25個與微藻生物膜量有關的光譜特征變量,計算出了每個特征變量的相關性系數(shù)。(5)采用相關性系數(shù)絕對值大小來衡量,提取出了與微藻生物膜干重關聯(lián)度前四的光譜特征變量,分別為:SDy)+SDy)/(SDr-(SDr、SDr、OSAVI、NDVI。將這四個光譜特征變量構建了三種數(shù)學預測模型,對比分析了BP神經(jīng)網(wǎng)絡預測模型、單一光譜特征變量以及多特征變量融合的預測模型,通過模型預測的耗時以及精確度來衡量,結果表明多特征變量融合模型對微藻生物膜量的綜合評價最高。(6)根據(jù)原始光譜反射值的變化以及微藻生物膜量的預測模型結果分析,小球藻生物膜培養(yǎng)的最佳生長環(huán)境是:溫度為28℃、PH值為8、光照強度為3500lx,而且預測模型還可以預測出不同時間的微藻生物膜量。
[Abstract]:The efficient cultivation of microalgae is the premise and key of bioenergy development and utilization. At present, the methods of microalgae culture include suspension culture and immobilized culture. Although the suspension culture model of microalgae is easy to operate, it is difficult to harvest, complex and the cost of harvesting is relatively high, while the immobilized culture model of microalgae has the advantages of stable operation, high biomass density, convenient and fast harvesting, low energy consumption, and so on. At present, microalgae researchers have become the focus of attention. Although the immobilized culture of microalgae has many advantages, the quality of microalgae, especially the biomass of microalgae, is significantly affected by biofilm growth factors. Therefore, it is very important to develop a fast and simple on-line nondestructive detection method to monitor the growth information of microalgae biofilm in order to adjust the growth conditions in time to improve the biofilm quantity of microalgae. As a new nondestructive, fast and accurate detection technology, hyperspectral imaging technology combines the advantages of computer vision and spectral detection, and can well record the rich information of biofilm growth. A simple and effective method can be found for rapid nondestructive detection of microalgae biomass. In this paper, the characteristics of microalgae biofilm reflectance spectrum and biofilm quantity prediction under different growing environment were studied and analyzed by combining hyperspectral technique with biological measurement technology. A hyperspectral monitoring method for microalgae biofilm is proposed, which provides a reference for the rapid acquisition of microalgae biofilm growth information and the efficient cultivation of microalgae biofilm. The results obtained are as follows: (1) the microalgae biofilm was cultivated at different temperatures, different PH values and different light intensities. (2) based on the theoretical basis of the hyperspectral imager, a hyperspectral image acquisition system was built. The hyperspectral information of microalgae biofilm under different temperature and light intensity was obtained by detecting the growth information of microalgae biofilm. There were three distinct characteristic peaks according to the original spectral reflectance of microalgae biofilm. The corresponding spectral reflectance at the characteristic wavelength of 789 nm ~ 811 nm ~ 930 nm decreased with the increase of culture time. (4) according to the spectral reflection data collected from microalgae biofilm, the spectral characteristic variables of biomass were measured. Twenty-five spectral characteristic variables related to biofilm amount of microalgae were extracted, and the correlation coefficient of each characteristic variable was calculated. The spectral characteristic variables of the first four levels of correlation with dry weight of microalgae biofilm were extracted, which were: 1 SDY) SDY / SDrSDr SSAVI / NDVI. The four spectral characteristic variables are used to construct three mathematical prediction models, and the BP neural network prediction model, the single spectral characteristic variable and the multiple characteristic variable fusion prediction model are compared and analyzed. The results show that the multivariate fusion model has the highest comprehensive evaluation of biofilm quantity of microalgae, according to the change of original spectral reflectance and the result of prediction model of biofilm quantity of microalgae. The optimum growth environment for the biofilm culture of Chlorella vulgaris was as follows: the temperature was 28 鈩，

本文編號：1897859