Verification Technology Of Affordable House Based On GF-1 Panchromatic Image

In our country,the affordable house was constructed to promote the “Anju Project” has played a significant role in solving the housing problems of low-and middle-income residents.The affordable house located in a remote rural area has a lot of phenomena such as "disorderly built,less built,and unbuilt" during the construction process.Resulting in waste and misappropriation of government finances.Therefore,in order to implement the government's investment,it is necessary to verify all the affordable house in these areas.Due to the vastness of the rural area,and the complicated geographical environment,it is difficult to collect information.Field-by-field verification is not only time-consuming and laborious,but also has a lot of unnecessary economic expenditure,that resulting in lower verification efficiency.If the current method which sampling field verification is adopted,the small proportion of sampling will easily lead to hidden errors.Therefore,due to solve those problems,the GF-1 panchromatic image will be used as basic data,combining the technology of remote sensing monitoring and remote sensing image processing,based on building feature extraction and segmentation algorithm to carry out research on how to achieve the method of fast and accurate verification of affordable house.The main research contents of this paper are as follows:(1)Research on building feature extraction algorithm of Morphological Building Index(MBI).By combining the features(grayscale,shape,and texture)of the building on the high-resolution panchromatic remote sensing image,the linear structural element with multiple scales and multiple directions is used to process the image by mathematical morphological and calculate the MBI characteristic map.After that,The conclusions and results are obtained by using multiple different parameters of MBI for image processing and grayscale feature analysis: The structural elements of 4 directions and 8 directions have no obvious influence on the algorithm,but are more sensitive to the scale factors of it;Finally,the algorithm can effectively extract building information when the setting parameters are reasonable.(2)An image segmentation algorithm of adaptive parameter Mean Shift and verification method of affordable house are proposed.The algorithm of adaptive parameter Mean Shift is improved based on the algorithm of traditional Mean Shift.On the one hand,the LUV(L is brightness,U and V are chroma)feature space is replaced by the MBI feature map,and the MBI feature map is masked that based on the vector data of the affordable house to filter out other interferences and highlight useful information.On the other hand,the traditional fixed-parameter method is abandoned,and the concept of standard deviation and gradient value which is calculated by the first-order partial derivative finite difference is introduced to modify the bandwidth of chromaticity and the parameter of segmentation in the Mean Shift algorithm to achieve Adaptive segmentation.The verification method of affordable house is based on the result of building extraction and the overlap rate with vector data.By checking the overlap rate within the range of the set threshold,verify the condition of each affordable house in each vector data.The experiment and quantitative evaluation index on the two-phase GF-1 panchromatic image show that the adaptive parameter Mean Shift algorithm in this paper has higher building extraction accuracy,better universality and robustness by comparing with the other three algorithms of image segmentation(threshold segmentation,Iterative Self-organizing Data Analysis clustering,traditional Mean Shift).It also verifies the effectiveness of the verification method of affordable house which is proposed in this paper.