I am having IFSAR DSM. I want to detect anomaly (both Spikes and depressions) to fix in the process of DTM editing. Since the file is too large I chose automation for the process of detection. I tried with custom kernel 3*3 sliding window to get the difference in elevation with comparing surrounding pixels. I categorized spikes (more than 10m) and depressions (less than -10m) based on the height difference. The spikes were good but the depressions were not plotted correctly. Is there any way to detect these kind of DSM anomaly using Erdas Imagine. I have Erdas Imagine 9.2, Global Mapper 13, ArcMap 10 and QGIS. Kindly impart your knowledge.