Massive Space database implementation policy-raster data 5

I. Some Influencing Factors

· Compression format and compression ratio

When importing raster data, you can select different compression formats and compression ratios based on your needs. Common examples include compression, lz77, JPEG, and MPEG-4, such as lz77 lossless compression, there are also JPEG and other lossy compression formats. You can also choose different compression quality for lossy compression formats. The following is a simple comparison of the storage and quality of raster data in different compression formats and compression ratios.

Use a gb tiff format without compressing raster data without pyramid as the data source and export it into several data with different compression formats and compression ratios. The results are as follows:

Compression format/compression ratio	Data Volume	Compression duration	A small range of preview time
Tiff/no compression	4.72 GB		2.2 seconds
PNG/lz77	3.92 GB	16 minutes and 8 seconds	1093.2 seconds
JPG/100%	2.05 GB	3 minutes 43 seconds	1473.0 seconds
JPG/75%	598 m	2 minutes 51 seconds	870.7 seconds
JPG/50%	396 m	2 minutes 26 seconds	827.7 seconds
Managed filegdb/uncompressed	4.76 GB	16 minutes 56 seconds	7.9 seconds
Managed filegdb/jpg/75%	1.73 GB	34 minutes and 6 seconds	20.3 seconds
ArcSDE/no compression	4.86 GB	41 minutes and 2 seconds	77.6 seconds
ArcSDE/. jpg/75%	1.72 GB	14 minutes and 11 seconds	20.3 seconds

By analyzing this result, we can draw the following conclusions:

1. It is very efficient to store large raster data without compressed files.

2. reading efficiency of big raster data using lz77, JPEG, and other compressed file storage is very poor, and data computing consumes too much resources.

3. lz77 compression algorithm is very limited and is not recommended (although it is the default value)

4. Using a JPEG compression algorithm with a compression quality of 75% is a good balance point, with a large compression volume and low image loss.

5. The difference in compression time between different quality using JPEG compression algorithms is not big.

6. File geodatabase is used to store large raster data. Even if JPEG compression is used, the reading efficiency will not be greatly reduced, but the non-compression performance is better.

7. Instead, ArcSDE provides better compression and storage performance than data without compression. It can be seen that the biggest factor affecting the performance of the database storage grid is the amount of data read.