Accurate information on change types is widely used in areas such as territorial spatial planning and dynamic monitoring of land cover, etc. Change type information is usually obtained using the post-classification comparison method. However, such methods are susceptible to the accumulation of errors due to the classification results, which in turn affects the accuracy of change type information identification. In response to the above problems, a change type determination method based on knowledge of spectral changes in land cover types (KSC-LCT) is proposed. Firstly, the change vector analysis method is used to obtain the change magnitude map, and the threshold value is calculated using the OTSU method to obtain the change regions. Secondly, some samples of land cover types are selected to calculate the mean value of each land cover type, and the spectral knowledge information of different land cover type changes is carved from the perspectives of spectral mean change, spectral angle change and thematic index change. Finally, based on the classification data of the previous moment and spectral change knowledge information, knowledge matching is carried out with the spectral change information of remote sensing images in the changed areas, and the minimum value of the matching result is selected as the judgment criterion to obtain the final change type of land cover. The experimental results show that the KSC-LCT method can effectively improve the recognition accuracy of land cover change type information and further reduce the phenomenon of over-detection in the change areas.

准确的变化类型信息广泛应用于国土空间规划、土地覆盖动态监测等领域，变化类型信息通常采用分类后比较法获取。但此类方法容易因分类结果产生误差累积，进而影响变化类型信息识别的准确性。针对上述问题，提出了一种基于土地覆盖类型光谱变化知识的变化类型判定方法（KSC-LCT）。首先利用变化向量分析法得到变化幅度图，利用OTSU法计算阈值得到变化区域。其次，选取部分土地覆盖类型样本，计算各土地覆盖类型的均值，从光谱均值变化、光谱角度变化和专题指数变化等角度刻画不同地表覆盖类型变化的光谱知识信息。最后，基于前一时刻的分类数据和光谱变化知识信息，与变化区域的遥感影像光谱变化信息进行知识匹配，选取匹配结果的最小值作为判断标准得到最终的土地覆盖变化类型。实验结果表明，KSC-LCT方法可以有效提高土地覆盖变化类型信息的识别精度，进一步减少变化区域的过检测现象。