The S-300 (NATO reporting name SA-10 Grumble) is a series of long-range surface-to-air missile systems developed by the former Soviet Union. It was produced by NPO Almaz for the Soviet Air Defence Forces to defend against air raids and cruise missiles.

2022年2月27日 · Detailing the technical specifications, development, and operational history of the SA-12 (Gladiator / Giant) / S-300V Long-Range, High-Altitude Self-Propelled Surface-to-Air Missile (SAM) System including pictures.

如苏联刚刚服役了一型最先进的地空导弹，西方赶紧给它起了个绰号，叫“斗士”，并规定“斗士”导弹的代号为SA-12。 萨姆系列 防空导弹 由前 苏联 1948年开始研制，由拉沃契金地对空导弹设计局在德国 “瀑布”地对空导弹 的基础上发展起来的，最早的型号是 萨姆-1防空导弹。 萨姆是 北约 给前苏联导弹起的代号，而前苏联命名时使用字母C，而俄语字母C对应英语字母S，萨姆（SA）是SANM的缩写。 虽然北约给前苏联起的绰号有些比较繁琐，而且改名、别称频繁，但“萨姆”这 …

The S-300VM " Antey-2500 " (Russian: С-300ВМ Антеӣ-2500, NATO reporting name SA-23 Gladiator/Giant) is a Russian anti-ballistic missile system. The system is designed to target short- and medium-range ballistic missiles, aeroballistic missiles, cruise missiles, fixed-wing aircraft, loitering ECM platforms, and precision-guided munitions.

2001年6月1日 · The S-300 series comprises the SA-10, SA-12, and SA-20 missiles and attendant radars. Each missile-radar combination is geared to operations within a range of altitudes and targets. It is the definitive “double-digit SAM” threat that has spurred the development of US stealth systems over the last 20 years.

2016年6月10日 · number one simultaneously fired SAM missiles: 12 : Probability of hitting one 9M83 missiles "Lance" - 0.65; Probability of hitting one aircraft: up to 0.9; Reload time : Warhead: 150 kg, HE:

2021年10月20日 · The first generation of PVO-SV area defence SAMs, the long range 2K11 / SA-4 GANEF and short range 2K12 / SA-6 GAINFUL were powered by ramjets, sacrificing missile speed and altitude performance to avoid the operational difficulties inherent in the higher energy liquid rocket propellants used by the Voyska PVO.

2024年8月20日 · Segment Anything Model 2（SAM 2）是由Meta公司发布的一个先进的图像和视频分割模型。它是Segment Anything Model（SAM）的升级版本，SAM是Meta的FAIR实验室发布的一款用于图像分割的基础模型，能够在给定提示的情况下生成高质量的对象mask。

stamp verification数据集，用sam预训练模型效果如下，虽然可以分割，但是不是非常完美，就需要微调提高分割精细度。该数据集标注真值包含mask和box，所以用box作为prompt输入进行微调。

2025年1月2日 · 可以用 from pycocotools import mask 这个模块解析。 下面是一段简单的可视化SA-1B数据集的masks的代码。 可视化效果如下. SA-1B数据集： https://ai.facebook.com/datasets/segment-anything/SAM公开了这个SA-1B数据集，包含1千万张图片和11亿个掩码，数据被划分为1100个tar压缩包，每个10G。 图片和标注文件（json）在同一 …