Testing the method when the reference image has time disparities

The effect of disparities in the time of acquisition between the high-resolution and reference images is a critical issue in cover types that are likely to have rapid anthropogenic, seasonal or natural change in reflectance character.

This has been tested with 8 DMSV images over varied topography using 4 Landsat TM images acquired at different dates and seasons over a range of cover types. The results showed that where there are differences between the Landsat TM images and DMSV the results are affected.

The following images were acquired over varied terrain in the Northern Territory. Figure 24, Figure 28, Figure 32 and Figure 36 were over woodlands, an irrigated area and a wetland. Figure 25, Figure 29, Figure 33 and Figure 37 were over an area dominated by a mine pit, rock dumps and woodlands. Figure 26, Figure 30, Figure 34 and Figure 38 were acquired over wetland areas. Figure 27, Figure 31, Figure 35 and Figure 39 were acquired over areas dominated by woodlands and mine structures. The images on the left show the raw DMSV images, the centre images are Landsat TM subsets used to model the brightness surfaces and the images on the rightside of the page are the DMSV images after the surfaces are removed.

The reference Landsat TM images acquired at time intervals of one year before (same season); one season before (different season dry versus wet vegetation condition); almost concurrent; and, two years after the DMSV acquisition. These were used to test the effects of time disparities on the correction.

Figure 24: LEFT: Two uncorrected DMSV images, rh34 and rh35 acquired on 28.09.95 for an area dominated by natural vegetation (note that the vegetation had changed because of recent flood irrigation). CENTRE: Landsat TM subset acquired on 23.08.94. RIGHT: DMSV images on leftcorrected with Landsat TM acquired 1 year earlier.

Figure 25: LEFT: Two uncorrected DMSV images, rh58 and rh59 acquired on 28.09.95 for an area dominated by mining infrastructure. CENTRE: Landsat TM subset acquired on 23.08.94. RIGHT: DMSV images on left corrected with Landsat TM acquired 1 year earlier.

Figure 26: LEFT: Two uncorrected DMSV images, rh68 and rh69 acquired on 28.09.95 dominated by varying cover types. CENTRE: Landsat TM subset acquired on 23.08.94. RIGHT: DMSV images on left corrected with Landsat TM acquired 1 year earlier.

Figure 27: LEFT: Two uncorrected DMSV images, rh78 and rh79 acquired on 28.09.95 with fairly uniform cover types. CENTRE: Landsat TM subset acquired on 23.08.94. RIGHT: DMSV images on left corrected with Landsat TM acquired 1 year earlier.

Figure 28: LEFT: Two uncorrected DMSV images, rh34 and rh35 acquired on 28.09.95 dominated by native vegetation but showing the effects of flood irrigation and a small fire. CENTRE: Landsat TM subset acquired on 22.05.95 showing evidence of the fire. RIGHT: DMSV images on left corrected with Landsat TM showing some contribution of the fire and flooding effects about 4 months apart.

Figure 29: LEFT: Two uncorrected DMSV images, rh58 and rh59 acquired on 28.09.95 for an area dominated by mining infrastructure. CENTRE: Landsat TM subset acquired on 22.05.95. RIGHT: DMSV images on left corrected with Landsat TM for a different seasonal condition.

Figure 30: LEFT: Two uncorrected DMSV images, rh68 and rh69 acquired on 28.09.95 with varying cover types. CENTRE: Landsat TM subset acquired on 22.05.95. RIGHT: DMSV images on left corrected with Landsat TM for a different seasonal condition.

Figure 31: LEFT: Two uncorrected DMSV images, rh78 and rh79 acquired on 28.09.95 with fairly uniform cover types. MIDDLE: Landsat TM subset acquired on 22.05.95. RIGHT: DMSV images on left corrected with Landsat TM for a different seasonal condition.

Figure 32: LEFT: Two uncorrected DMSV images, rh34 and rh35 acquired on 28.09.95 dominated by native vegetation and the flood irrigation. CENTRE: Landsat TM subset acquired on 27.09.95. RIGHT: DMSV images on left corrected with Landsat TM acquired 1 day apart.

Figure 33: LEFT: Two uncorrected DMSV images, rh58 and rh59 acquired on 28.09.95 for an area dominated by mining infrastructure. CENTRE: Landsat TM subset acquired on 27.09.95. LEFT: DMSV images on left corrected with Landsat TM acquired 1 day apart.

Figure 34: LEFT: Two uncorrected DMSV images, rh68 and rh69 acquired on 28.09.95 with varying cover types. CENTRE: Landsat TM subset acquired on 27.09.95. RIGHT: DMSV images on left corrected with Landsat TM acquired 1 day apart.

Figure 35: LEFT: Two uncorrected DMSV images, rh78 and rh79 acquired on 28.09.95 with fairly uniform cover types. CENTRE: Landsat TM subset acquired on 27.09.95. RIGHT: DMSV images on left corrected with Landsat TM acquired 1 day apart.

Figure 36: LEFT: Two uncorrected DMSV images, rh34 and rh35 acquired on 28.09.95 95 dominated by native vegetation but showing the effects of flood irrigation and a small fire. CENTRE: Landsat TM subset acquired on 31.07.97. RIGHT: DMSV images on left corrected with Landsat TM subset in acquired two years later.

Figure 37: LEFT: Two uncorrected DMSV images, rh58 and rh59 acquired on 28.09.95 for an area dominated by mining infrastructure. CENTRE: Landsat TM subset acquired on 31.07.97. RIGHT: DMSV images on left corrected with Landsat TM acquired two years later.

Figure 38: LEFT: Two uncorrected DMSV images, rh68 and rh69 acquired on 28.09.95 with varying cover types. CENTRE: Landsat TM subset acquired on 31.07.97. RIGHT: DMSV images on left corrected with Landsat TM acquired two years later.

Figure 39: LEFT: Two uncorrected DMSV images, rh78 and rh79 acquired on 28.09.95 with fairly uniform cover types. CENTRE: Landsat TM subset acquired on 31.07.97. RIGHT: DMSV images on left corrected with Landsat TM acquired two years later.