To realize the transmission of fuze radome deformation images under low bandwidth condition of the missile-borne telemetry radio， an image compression algorithm with high-quality compressed images， low algorithmic complexity and easy to implement in hardware was proposed. The 9/7M wavelet basis recommended by the Consultative Committee for Space Data Systems （CCSDS） was used to decompose the image at four levels， thereby reducing data redundancy. The traditional weight allocation strategy was improved according to the image characteristics to reduce the distortion information， ensuring the quality of the reconstructed image. A parallel pixel scanning approach was adopted， which significantly reduced the time associated with the multi-resolution tree-structured partitioning in Set Partitioning In Hierarchical Trees （SPIHT）. Furthermore， for the specific features of the images， an enhanced Run-Length Encoding （RLE） algorithm was introduced， the image data classification method was optimized， achieving increased overall compression ratios without compromising image quality. Experimental results show that in comparison with SPIHT and CCSDS standards， the proposed algorithm achieves 6.92%， 16.45% and 11.12%， 22.56% increases in Structural Similarity Index （SSIM） and Peak Signal-to-Noise Ratio （PSNR）， respectively， while the compression time increases by only 0.06 s compared with SPIHT， and the increase in algorithmic complexity is not much. Compared with the mainstream image compression algorithms， under the same compression multiplier， this algorithm compresses the fuze radome deformation images with similar compressed image quality as the JPEG2000 algorithm， but the compression time saving is up to 35.35%. In summary， the algorithm has the characteristics of low complexity， its compression performance can meet the compression requirements of the ballistic system， and the decompression image quality reaches the level of mainstream algorithms.