考虑局部自相似性的土地复垦规划空间预测算法

Land reclamation planning spatial prediction algorithm considering local self-similarity

土地复垦规划时土地图像存在较大噪声,且不同地区的土地复垦情况可能存在空间异质性,无法达到理想空间预测效果。为此,提出一种考虑局部自相似性的土地复垦规划空间预测算法。利用低秩逼近处理方法对噪声污染的观测矩阵的核范数低秩逼近,得到矩阵的最小核范数,实现土地空间图像去噪;利用局部自相似性计算土地再利用贡献值,以此建立空间预测模型,计算子区域内栅格与所选区域栅格差异程度;利用ArcGIS软件计算评价因子,以此划分等级,预测土地复垦规划空间。由实验结果可知,该算法预测的空间规划基本适宜度比例最大,其次是不适宜度、中度适宜度、临界适宜度,与理想规划结果一致。

There is significant noise in land images during land reclamation planning,and there may be spatial heterogeneity in land reclamation situations in different regions,which cannot achieve ideal spatial prediction results.Therefore,a land reclamation planning spatial prediction algorithm considering local self-similarity is proposed.By means of low rank approximation processing methods,the kernel norm of the observation matrix of noise pollution is approached with low rank approximation,and the minimum kernel norm of the matrix is obtained to realize denoising of land spatial images.The local self-similarity is used to calculate the contribution value of land reuse,so as to establish a spatial prediction model,and calculate the grid difference between the sub region and the selected region.ArcGIS software is used to calculate the evaluation factors,so as to classify the grades and predict the land reclamation planning space.The experimental results show that the algorithm can predict the highest proportion of basic suitability for spatial planning,followed by unsuitability,moderate suitability,and critical suitability,which are consistent with the ideal planning results.