Аннотация
This article presents CSTM (Compression-Structure Targeted Multispectral image’s Archiving Method), a specialized binary archive format for image collections with embedded metadata. Beyond multispectral imaging, CSTM targets intelligent mobile systems and Internet of Things (IoT) deployments where devices and edge nodes capture and relay large volumes of images under strict bandwidth, latency, and energy constraints. Universal archivers such as ZIP and RAR add 5–15% structural overhead and lack native image metadata support, which becomes a bottleneck in mobile/edge pipelines. The CSTM algorithm employs a minimal-istic fat binary structure with embedded metadata. Each image adds only 20 bytes of overhead plus filename length (compared to 100–300+ bytes in conventional formats). Optional resizing and format optimization precede unified zlib compression that leverages inter-image redundancy. On a set of 10 JPEG images totaling 24 MB, CSTM produced an 18.1 MB archive in 1.2 s versus ZIP’s 19.8 MB in 1.8 s and RAR’s 19.5 MB in 2.4 s—an 8–10% size improvement with 25–35% faster processing. In IoT/mobile pipelines this translates into shorter transfer times, reduced radio airtime, and faster edge/cloud ingestion.