SATELLITE IMAGERY OR ORTHOPHOTOS?

Mike Tuohy

New Zealand Centre for Precision Agriculture, Massey University

Remote sensing is becoming an important ingredient in the mix of precision agriculture. The interpretation of aerial photographs has been a particularly useful skill for many of those who have an interest in the land and its ability to grow things. Today remotely-sensed data comes in many forms: as well as the many types of aerial photographs, other airborne sensors generate information from various parts of the electromagnetic spectrum and satellites beam down thousands of images every day.

How do you chose which imagery is best for your purpose? Of course this will depend on whether you want to estimate carbon credits in a forest or monitor weed infestation in a five-hectare paddock. Mapping soils over a whole farm requires imagery at a different scale to that necessary for determining possible nutrient deficiencies in a maize crop. What about identifying poorly-producing trees in an apple orchard, or moisture stress in a vineyard, or blight in a potato crop?

LINZ has recently made 1:25000 digital orthophotos freely available on their website. Each image covers approximately a quarter of a NZMS 260 1:50000 topographic map. Two formats are provided: jpg files are usually less than 100 Kb and have a 25m spatial resolution, tif files can be as large as 50 Mb but have a correspondingly higher resolution - 2.5m. The only problem is that they are only greyscale images.

Some organizations have commissioned special aerial surveys for the acquisition of orthophotos. For example, Massey University has orthophotos of its campuses with spatial resolutions of 400 and 125mm. The higher resolution image of the Palmerston North campus is a 215 Mb file! Once again these are greyscale photos.

The most useful and readily available satellite images are those from Landsat or SPOT. SPOT is a French system which has two main image formats: multispectral (green, red and near-infrared) at 20m spatial resolution and panchromatic (greyscale) at 10m resolution. The latest in the Landsat series was launched in April 1999 and is now providing very good imagery when weather conditions permit. Landsat 7 imagery has eight bands with spatial resolutions of 60m (thermal infrared), 30m (R,G,B,near and mid-infrared) and 10m (panchromatic).

Examples of all these images will be displayed and the relative merits for each, as well as suitable applications, will be discussed.

Return to Paper List