TEXTURE MEASURES FOR THE IDENTIFICATION AND MONITORING OF URBAN DERELICT LAND

被引:10
|
作者
DAWSON, BRP [1 ]
PARSONS, AJ [1 ]
机构
[1] UNIV KEELE,DEPT GEOG,KEELE ST5 5BG,STAFFS,ENGLAND
关键词
Fractal roughness - Run length - Standard deviation - Texture measures - Thermal band imagery - Urban derelict land;
D O I
10.1080/01431169408954159
中图分类号
TP7 [遥感技术];
学科分类号
081102 ; 0816 ; 081602 ; 083002 ; 1404 ;
摘要
Results are presented from a study in which texture measures are employed as a means of detecting urban dereliction, using the Potteries conurbation as a test site. Guided by the experience of using aerial photography for this task and also by a consideration of the workings of the human visual system, four texture measures. namely, standard deviation, entropy, run lengths and 'fractal' roughness, were selected for use with thermal band imagery. Standard deviation images show derelict sites to be characterised by frequent, low-intensity boundaries whereas rural areas have few boundaries picked out and those of the urban area arc high-intensity. Such images are potentially useful tools for monitoring the extent of derelict sites. Entropy images depict derelict land in mid-grey tones, between the high values characteristic of rural areas and the low values of urban areas. Run length images are perhaps the most successful at segmenting the derelict sites from the background but are most effective on large sites. The 'fractal' roughness image bears a strong resemblance to the standard deviation measure. These are promising results, but futher work is necessary to integrate them into a classification methodology.
引用
收藏
页码:1259 / 1271
页数:13
相关论文
共 50 条