Land use and its landscape pattern configuration are closely related to urban development, and Land Use/ Cover Change (LUCC) profoundly impacts regional soil and water conservation and ecological security.The spatio-temporal change characteristics of land use and landscape patterns in Zhongshan City, which is an important city in the Guangdong-Hong Kong-Macao Greater Bay Area, were explored at multiple scales and levels using analytical methods such as transfer matrix and landscape pattern index based on the land use data of GlobeLand30 from 2000 to 2020. The result shows:①In the past 20 years, the overall land use in Zhongshan City showed a trend of “two increases and three decreases” in which construction land and forest land increased, and cultivated land, grassland, and water bodies decreased. Construction land has grown rapidly, while cultivated land has plummeted by nearly 50%.②In the past 20 years, each land use type has been converted to each other, which is mainly manifested in the conversion of cultivated land, water bodies, and grasslands into construction land.③The dominant landscape type of Zhongshan City has changed from cultivated land in 2000 (670.22 km2, 37.23%), water bodies in 2010 (505.15 km2, 28.06%), to construction land in 2020 (644.43 km2, 35.80%), and urban development occupies a large amount of cultivated land and water bodies. At the same time, the scope of the high-value areas of the Patch Density（PD） and Contagion index（CONTAG） of construction land gradually increased, the degree of agglomeration was enhanced, and the landscape advantages continued to expand. In the past 20 years, Zhongshan City has shown a homogenization trend of land use types with construction land replacing other land types, and the spatial heterogeneity of the landscape pattern has gradually diminished, making the task of urban cultivated land protection and ecological construction more and more severe.

In order to identify the volatile flavor compounds from normal and crispy Nile tilapia, these components were confirmed by solid phase microextraction-gas chromatography-mass spectrometry (SPME-GC-MS). The results showed that the volatile flavor compounds from these two tilapias were different, and the main volatile flavor compounds from two Nile tilapia were heptanal, octanal, decanal, 2,4-decdienal, nonaldehyde, 1-octene-3-ol and 2,3-octanedione. The content of the main volatile flavor compounds of crispy Nile tilapia such as heptanal, octanal, nonanal, 1-Octen-3-ol is more than that in normal Nile tilapia, while contents of decanal and 2,3-octanedione are less than that in normal Nile tilapia, and there is no significant difference in 2,4-decadienal. The composition and concentration of volatile flavor compounds from crispy Nile tilapia were various with different heat treatments. While, heat treatment on two Nile tilapia can reduce content of the fishy smell compounds and change the chemical composition of volatile compounds. Among them, the roasting process can increase butter flavor and produce new lemon flavor, giving the fish a special flavor. Exploring the effects of two Nile tilapia and different heat treatments on the volatile flavor compounds can provide some theoretical guidance for food processing and storage technology, A theoretical basis and related data for the development of crispy tilapia prefabricated dishes, processing technology selection and product follow-up.

In order to evaluate the effect of storage periods on the quality of compound feed for soft-shelled turtle (Pelodiscus sinensis) at room temperature, the feed samples were collected at 0, 7, 14, 28, 35, 45 and 60 d, respectively. The feed quality was evaluated by the indexes, including moisture, crude protein, crude fat and volatile salt-based nitrogen (TVB-N), and their changes with storage periods were analyzed, so as to determine the optimal storage period of compound feed for Pelodiscus sinensis. The results showed that there was no significant difference in crude protein content among the groups within 0-35 d of storage. However, the crude fat content decreased significantly after 28 d compared with that of the control group (0 d). The moisture in the diet of soft-shelled turtle decreased at first and then increased after 28 d of storage, which was significantly higher than that of the control group(0 d). The content of volatile basic nitrogen increased rapidly after 28 d of storage, and the difference was significant. According to the standard of sensory index, the appearance of the feed was normal within 45 d after storage under this experimental conditions. The smell of the feed changed, and the smell gradually faded with a musty smell from 35 d. The results showed that the quality of feed for Chinese soft-shelled turtle did not change significantly during storage 0-28 d. This study provided reference data for the storage of Chinese soft-shelled turtle formula feed.

Honey and sucrose polyester were used to partially replace traditional syrup and peanut oil respectively. The formula of low sugar and fat cantonese moon cake skin was optimized by Complete randomalized design and Orthogonal experimental design. Combined with the value of texture profile analysis and sensory evaluation index, it was comprehensively analyzed. The experimental results showed that the best formula of low sugar and fat cantonese moon cake crust was: 100 g flour, 50 g conversion syrup and 20 g peanut oil as base, honey substitution rate of 15%, Jianshui amount of 3 g and sucrose polyester substitution rate of 15%. The moon cake crust obtained under this condition had complete shape, soft taste, non greasy and no sticky feeling.

In order to reduce the supercooling degree of disodium hydrogen phosphate dodecahydrate and improve its water loss problem,a composite phase change material (CPCM) was prepared by adsorbing disodium hydrogen phosphate dodecahydrate with SiO2 as carrier material in this paper. The effect of three nucleating agents on the supercooling degree of the CPCM was investigated through T-history method. The chemical structure and thermal properties of the CPCM were characterized. The results showed that using 5% sodium silicate nonahydrate as a nucleating agent could reduce the supercooling degree of the CPCM to 2.3 ℃. The Differential Scanning Calorimeter (DSC) results displayed that adding a certain amount of deionized water could effectively improve the water loss problem of the CPCM, with its phase transition temperature and latent heat were 35.1 ℃ and 152.5 J/g，respectively. The Fourier Infrared Spectrometer (FT-IR) results showed that there was only physical adsorption between SiO2 and disodium hydrogen phosphate dodecahydrate. The Thermo Gravimetric Analyzer (TG) results indicated that the CPCM had good thermal stability.

Cirrus clouds were widely existed in the atmosphere, which seriously hindered the ground observation process of optical remote sensing, and reduced the accuracy and availability of data. An automatic cirrus clouds correction method was proposed, which fully uses the cirrus clouds detection band and couples a scattering law with the band correlation. Firstly, a scattering law was used to quantitatively describe the cirrus clouds intensity and further applied to derive the correlation of cirrus clouds in bands. Then, taking the cirrrus clouds detection band as reference, the band correlation between adjacent bands was mined and coupled with a scattering law to estimate the cirrus clouds intensity accurately in different bands and remove the cirrus clouds. The Landsat-8 operational land imager (OLI) data was selected for experiment to validate the accuracy and adaptability of the method. Moreover, the airborne visible infrared imaging spectrometer (AVIRIS)and GaoFen-5 advanced hyperspectral imager (AHSI)data were further chosen to test the expandation. Results showed that the proposed method could effectively remove the cirrus clouds interference in different bands and was suitable for various scenes, accurately recovered the reduced surface information, had good performance in visual and quantitative evaluation, and could meet the application requirements of quantitative remote sensing.

In order to solve the adverse impact of river and lake sludge on the environment，a study was carried out on preparing of permeable pavement bricks with river and lake sludge. The representative river and lake sludge in Guangzhou was selected，and the basic physical and chemical properties of the sludge were analyzed. The river and lake sludge was used to replace part of the concrete aggregate to prepare the base material. Then the tests were carried out with sludge content 0.0%～30.0%, the water cement ratio 0.30～0.38, and the glue bone ratio 1∶3.0～1∶4.5. The flexural resistance, permeability coefficient of permeable were determined bricks. Based on the analysis of the experimental data, the feasibility of preparing permeable pavement bricks by using river and lake sludge and the corresponding mix ratio were studied. The optimum basic ratio was obtained with glue-bone ratio 1∶3.0, sludge mixing amount 10.0%, water-cement ratio 0.32, which provided a reference for the resource utilization of river and lake sludge.

Banana is easily infected by spoilage fungus during transportation and post-harvesting storage. Identifying process of fungus infecting banana is conducive for timely detection of potentially infected fruit and taking scientific measures for prevention and control. In this paper, the near-infrared spectra (930-1 650 nm) of banana fruit infected by Fusarium solani at different infective stages were collected. Based on the full range data, principal component analysis-support vector machine (PCA-SVM) discriminant model and partial least squares discriminant (PLSDA) model based on the original spectrum were established respectively after comparing effects of different preprocessing methods on the models. Both models achieved good results with discriminant accuracies of 83.33% and 76.67% for validation sets, respectively. Furthermore, ten characteristic wavelengths (1 117.5,1 140.7,1 146.4,1 255.5,1 284.0,1 312.5,1 403.2,1 493.2,1 498.8,1 621.5 nm) were screened out using competitive adaptive reweighted sampling (CARS) algorithm, and SVM and PLSDA models were established based on these characteristic wavelengths, respectively. The performance of CARS-SVM model was better than that of CARS-PLSDA model, with identification accuracies of 84.78% and 78.57% for training and validation sets, respectively. Results indicated that NIR spectra could be used to identity process and degree of Fusarium solani infecting banana fruit.

Maturity judgement is an important basis for mango harvesting and storage. In order to meet the limitation of hardware computing capacity of mobile devices, the performance of traditional machine learning and transfer learning in mango maturity identification was compared. The optimal model was selected and the mango maturity classification software was developed. One hundred images of Xiaotainong mango in different maturity were collected and then divided into training set and test set with a ratio of 8∶2 after data expansion. Accuracy, F1 value and prediction time were used as model evaluation indexes. Four machine learning algorithm models (K-NearestNeighbor, Support Vector Machine, Naive Bayes and Decision Tree) and five transfer learning algorithm models (AlexNet, ResNet18, VGG16, GoogleNet, and SqueezeNet) were used to train and test respectively. The performance of each model was compared and analyzed. The results showed that although machine learning had strong computing ability, the classification accuracy of machine learning was significantly lower than that of transfer learning. The classification accuracy of all transfer learning was >90%. Considering both classification accuracy and computing power, resnet18 performs best with accuracy rate reaches 98.75% and test time is only 74.66 ms after 20 iterations, which is superior to other transfer learning models.

Divided differences are very important in numerical approximation and are introduced completely in different numerical approximations literatures. Divided differences with repeated nodes are defined on the basis of divided differences. However, they are not systematically summarized in different literatures. Some main properties of divided differences with repeated nodes are summarized, and their applications in the construction of the Hermit interpolation are also introduced and its theoretical basis has been proved.

Compared with traditional transplanting methods, direct seeding rice has gradually developed into an important planting method which is widely welcomed in China due to its no raising rice seedlings and transplanting steps, high production efficiency and low production cost. However, lodging is a key problem in the production of direct seeding rice, which often leads to rice yield reduction, even no harvest, and rice quality decrease. In production, nitrogen application rate and operational research are one of the key ways to improve lodging resistance of direct seeding rice. The lodging types of direct seeding rice, the effects of nitrogen on lodging of direct seeding rice, the nitrogen absorption characteristics of direct seeding rice, and the effects of optimizing nitrogen management measures on lodging resistance of direct seeding rice were reviewed, so as to provide reference for the construction of nitrogen regulation and cultivation techniques for lodging resistance of direct seeding rice.