Applications Of Ga-Svm Algorithm Of Land Use/Land Cover Classification And Research Of Land Use/Land Cover Dynamic Changes In Baojixia Irrigated

Baojixia Irrigation is the largest irrigation area, and is also the most important grain, oil, fruit and vegetable production base in shaanxi province. Therefore the paper uses types to classify and statistical analysis of Baoji gorge irrigation area based on remote sensing method, timely and accurate understanding the Land-Use and Land-Cover Change (LUCC) of Baojixia Irrigation. This analysis has important significance to the management planning, rational layout and reasonable distribution of irrigation water of irrigation area resources. Using the model to forecast the future land use conditions could provide theoretical basis for the government’s land management planning, in order to realize the sustainable utilization of the land.The paper is based on the phenological calendar in the study area and land use types, to analysis the remote sensing image, establishing interpretation signs, select the blueprint of each area. Use TM1-TM6and TM7altogether six band grey value and NDVI, NDWI and NDBI three dimensional matrix as the initial data of index value; Based on Matlab GUI development GA-SVM algorithm software for initial data optimization, get the optimal parameters y and C of support vector machine classifier, and then to remote sensing image classification; According to the changes of the study area in different periods of land use type, combined with statistical data to analysis the cause of land use changes in Baoji gorge irrigation area through three dimensions:natural factors, economic factors and social factors. Finally, use the Markov model to simulation prediction the change trend of agricultural land (land and garden land) of Baoji gorge irrigation area in the coming18years.In remote sensing image, for example, on January21,2009, the classification results show that the overall accuracy of genetic algorithm to optimize support vector machine classification is95.31%, kappa coefficient is0.9374, combined with a grid optimization, and also increased by5.63%and11.14%combined with the neural network classification and the minimum distance classification. The results show that genetic algorithm optimization of support vector machine classification is better than the traditional method of remote sensing classification, and it is an effective method to remote sensing classification of Baojixia Irrigation. Between1995and2009, garden land area is increased from364.03km2(19.68%) to672.26km2(36.35%), mainly because the irrigation area is developing the fruit in recent years, and part of the unused land turned into gardens. During the monitoring time, and continuously cultivated land turned into building land and garden land, the area is in decline, from the original879.00km2(47.52%) to759.79km2(41.08%). Water from17.21km2(0.93%) in1995fell to13.05km2(0.71%) in2009, the change is mainly caused by the river diversions and seasonal change.Economic factors and social factors are the main factors of land use type change of Baojixia Irrigation. Although the natural factors also has a certain guiding role in the development of Baoji gorge irrigation area, but its role is far less than the other two factors on the impact of land use change. Under the condition of invariable of human factors influence, the paper predicts the change trend of cultivated land and garden land of Baojixia Irrigation in2015,2021and2027by Markov model, the results show that the change trend of garden land and cultivated land in the coming18years is almost the same combined to the present stage. Garden land area is expanding under the drive of economic and policy factors, and cultivated land is decreasing due to the conversion to other land use types. The garden land area will increase to681.58km2,823.31km2and871.94km2in2027, respectively36.59%,44.25%and46.93%of the total area. And the cultivated land will in turn reduce to757.02km2,720.79km2and714.49km2, accounted for about40.64%,38.74%and38.46%of the total area. So that the government’s planning should be considered both social and economic development. In the meantime, make out a strong need for cultivated land protection, scientific development and utilization of land resources, further implementing land use control system, strict implementation of general land use planning.