左手亲情右手爱

Grassroots official’s 30-year fight against desertification in North China_我的网站

A |     7月22日10时54分，引力一号遥四运载火箭在我国东海海域点火升空，一箭将9颗卫星送入预定轨道。

B |     这9颗卫星包括东坡13星、14星、17星至20星，西光贰号01星、天仪49星和紫丁香三号，同时还开展了一项海王星载荷试验。    

    Wensuyalatu checks the growth of sour jujube seedlings. Photos: Chinanews.comAcross the vast desert landscape of the Urad Rear Banner in North China's Inner Mongolia Autonomous Region, one grassroots official has spent more than three decades turning barren sands into patches of green.
Ecological restoration
For over 30 years, Wensuyalatu, a member of the leadership group of the Urad Rear Banner forestry and grassland bureau, has made the Gobi Desert his workplace, dedicating his life to combating desertification, advancing ecological restoration through technological innovation, and helping local herders increase their incomes, according to the local bureau's official WeChat account.
Located on the edge of the desert in Bayannur's Urad Rear Banner, it is a key area in China's Three-North Shelterbelt Forest Program, a major ecological project aimed at strengthening the country's northern ecological security barrier. 
However, early afforestation efforts in the desert were far from easy at the start.
The project covers 103,800 mu (about 6,920 hectares), including comprehensive treatment of 24,500 mu of desertified land, ecological restoration of 4,100 mu of non-desertified land, and 75,200 mu of enclosed grassland restoration. The project also includes the construction of 11 small-scale water conservation facilities, all of which required strict standards and extensive coordination.
The project also includes the construction of 11 small-scale water conservation facilities, all of which required strict standards and extensive coordination.
In 1991, Wensuyalatu, then in his early 20s, chose to devote himself to forestry and grassland conservation in the Urad Rear Banner. Since then, he has remained on the front lines of desertification control, witnessing the region's harshest environmental challenges and working to find solutions adapted to the fragile desert ecosystem.
In the early years, afforestation in the Gobi Desert was an extremely difficult task. The most critical planting season in spring often coincided with the strongest sandstorms of the year. Every year during the key afforestation period, Wensuyalatu would stay deep in the desert, guiding workers on planting techniques and management practices, while carefully overseeing every detail of the ecological restoration projects.
"Every spring, the wind and blown sand were so strong that it was hard to keep your eyes open. The sand hit your face like needles. During the day, we followed workers to dig holes and plant trees, and at night we stayed in temporary shelters to organize data and review the day's work. That was the normal routine for desertification fighters," Wensuyalatu recalled.
The road to restoring the desert was far from easy. Years of drought, low survival rates of seedlings, and skepticism from some local residents created major obstacles.
"At the beginning, it was really difficult. The climate was extremely dry, and many seedlings died after being planted. Some herders did not understand or support the work. They believed deserts were naturally barren and that no amount of effort could make trees grow there," Wensuyalatu said.
To change these perceptions, Wensuyalatu brought local residents to see firsthand the damage caused by desert expansion. Through visible results and long-term communication, he gradually won the understanding and support of the community.
Facing the ecological challenges of the Urad Rear Banner, Wensuyalatu began exploring new approaches beyond traditional restoration methods.
    Staff members get ready to carry out aerial seeding operations.Aerial seeding
Since 2004, he has led efforts to tackle desert restoration challenges by introducing large-scale aerial seeding projects, creating a breakthrough approach for restoring degraded desert ecosystems in the region.
"Compared with traditional methods, aerial seeding is more efficient and can cover wider areas. Based on local climate and soil conditions, we select suitable desert plant seeds and precisely control the seeding height, density and timing to improve survival rates as much as possible and accelerate desert recovery," Wensuyalatu explained.
After years of effort, the aerial seeding projects under Wensuyalatu's leadership have restored more than 800,000 mu of desert land, with an effective restoration rate of 62.2 percent. 
He also developed an integrated management approach combining "protection, management, retreat, restoration and publicity," establishing a mixed vegetation system of trees, shrubs and grass to curb desert expansion and reduce sandstorms at their source.
Beyond restoring greenery, Wensuyalatu has also focused on strengthening scientific research and protecting native desert species.
Starting in 2010, he devoted himself to the artificial cultivation of endangered native desert plants. After more than a decade of repeated experiments, he successfully developed cultivation techniques for rare species such as the endangered evergreen broadleaf shrub Ammopiptanthus mongolicus, Mongolian almond and drought-resistant elm.
Under his leadership, his team cultivated 6.5 million seedlings and completed more than 4,000 mu of high-quality afforestation projects, with seedling survival rates consistently exceeding 80 percent. 
    Engineering personnel operate drones to conduct aerial seeding as part of afforestation efforts in the Badain Jaran Desert in North China on June 17, 2025. Photo: VCGHis achievements filled a technological gap in large-scale cultivation of native desert plants and earned him second prize in the 2017 Inner Mongolia Autonomous Region Harvest Award.
For Wensuyalatu, desert restoration is not only about creating ecological benefits, but also about improving people's livelihoods.
He has long promoted the idea that "trees should bring both ecological value and economic value." By leveraging local desert resources, he helped establish a 10,000-mu industrial base for desert plants and cistanche - a medicinal plant often grown in desert environments - across four key local villages, integrating ecological restoration with specialty desert industries.
"In the past, people thought desertification control was simply about planting trees - it took a lot of efforts but brought little income. Now things are different. By planting trees and developing industries such as cistanche cultivation, desert land can generate real wealth," a local herder said.
To ensure that local communities benefit from ecological development, Wensuyalatu introduced enterprises, supported family forest farms and professional cooperatives, and promoted large-scale, market-oriented operations.
The once barren desert has gradually become a "green treasure chest" for local herders, achieving a win-win outcome: greener landscapes, improved ecosystems, stronger businesses and higher incomes.
"Desertification control is not just an ecological project, but also a project that benefits people," Wensuyalatu said. 
"Only when communities gain real benefits from protecting nature will they actively participate in ecological conservation. That is the key to making restoration efforts sustainable and achieving the vision that lucid waters and lush mountains are invaluable assets," Wensuyalatu said. 
Global Times
。    这批卫星里，有SAR卫星，有光学遥感卫星，也有新技术试验星。    老颜觉得，中科西光自主研制的西光贰号01星，也就是彩云高光谱01星，蛮有看点的。    按照中科西光的说法，这是我国首颗定量高光谱智算卫星，也是中科西光与云南地矿合作的首颗卫星。         更引人注意的是，中科西光还把248 TOPS算力和行业AI模型一起装进了卫星。    遥感卫星原本主要负责拍片，数据传回地面以后再慢慢分析。    现在，“读片室”也被送上了太空。         太空算力，得先找一件具体的事做          太空算力讲到今天，最不缺的是宏大计划。

C |     轨道数据中心、太空大模型、成千上万颗算力卫星连成网络，每一个故事都足够震撼。    但真要做成，还得过发射成本、能源、散热和大规模组网几道关。    中科西光选择的路线比较直接。    数据既然在卫星上产生，就先在卫星上处理。    这不是把一座大型数据中心搬上太空，而是给遥感卫星增加足够的计算能力，让它拍完以后先做一轮识别和筛选。    西光贰号01星搭载了高光谱光学载荷、全色光学载荷和在轨智能处理载荷，设计运行于535公里太阳同步轨道。    全色相机主要看清地面的形状和轮廓。    高光谱载荷解决的是另一类问题：地面上的东西是什么，成分和状态有没有发生变化。    矿物、水体、森林和农作物，反射光线的方式并不相同，也就留下了各自的光谱“指纹”。    一块岩石里可能含有什么矿物，庄稼这一季长得怎么样，河流里的泥沙和藻类有没有明显变化，都可以从这些细微的光谱差异中寻找线索。    所以，高光谱经常被形容为给地球做CT。    普通卫星主要拍下一张图，高光谱卫星则要为图像中的每个像素记录一条光谱曲线。    信息更丰富，数据量也大得多。    这正是它需要星上计算的原因。

D |          这台“太空CT”，前端也很强          如果只盯着248 TOPS，很容易忽略这颗卫星前端的高光谱载荷。    算力决定它能处理多少数据，载荷则决定它拍回来的数据到底够不够细、够不够准。

E |     西光贰号01星采用新型静态干涉光谱成像技术，减少了传统光谱仪里的运动部件。    按照设计，这样更有利于卫星长期稳定运行。    中科西光公布的参数是10米空间分辨率、7纳米光谱分辨率、14公里幅宽，单星每天最高可以获取100万平方公里高光谱影像。    10米，决定它能看清多小的地面目标；7纳米，决定它能不能分辨性质接近的物质；14公里幅宽和每天100万平方公里，则决定它一天能干多少活。    实际成像还会受到云层、光照和任务安排影响，但这组参数至少说明，它不只追求看得细，还要兼顾大范围连续观测。    不过，真正让名字里多出“定量”两个字的，不只是这组参数。

F |     普通高光谱识别，更多是判断这里可能有什么。    定量遥感还要继续往下算：泥沙浓度有多高，水体浑浊到了什么程度，不同区域之间差了多少。    所以，这颗卫星不只想拍一张“太空CT”。    它还准备在轨道上先做一轮初步读片。         248 TOPS，放在全球是什么位置？          其实，星上处理高光谱数据，国外已经探索了二十多年。    NASA的EO-1卫星从2003年前后开始运行自主科学航天器软件，在轨分析高光谱和多光谱数据，还可以根据识别结果调整后续观测任务。    2020年，欧洲空间局发射PhiSat-1，利用星上AI识别云层，提前筛掉被云遮挡的图像，减少无效数据下传。

G |     2023年发射的波兰Intuition-1，与西光贰号01星的路线更为接近。    它搭载192波段高光谱传感器和3 TOPS星载处理单元，可以在轨开展图像分割、云层筛选和高光谱分类。    西光贰号01星公布的算力达到248 TOPS，纸面数字大约是Intuition-1的80倍。

H |     这个80倍不能直接换算成实际处理速度。    两套设备采用的芯片架构、计算精度、功耗和算法不同，TOPS的统计口径也未必一致。    但把248 TOPS装进一颗商业高光谱卫星，分量还是够的。    起码，中科西光不是全球第一个做星上高光谱处理的，却把这件事做到了更大的算力规模。    前面的海外项目已经证明，星上筛选、识别和自主调整任务是可行的。

I |     西光贰号01星继续往前走，把水体泥沙等行业模型也装了进去，准备判断河水哪里变浑、矿区哪里发生变化、农作物长势有没有异常。    这就不只是演示算法了，开始往具体业务里走。         卫星开始在天上“看懂”数据          西光贰号01星搭载了中科西光与之江实验室联合研发的高光谱水体泥沙分级模型。    卫星飞过一条河流时，可以先在轨判断不同河段的浑浊程度，标记值得关注的位置，再把分级结果和重点区域的光谱数据传回地面。    地面系统继续做精细分析和人工复核。    客户收到的，不再只有一大包等待解译的原始数据，还多了一份可以优先检查的目标名单。    这套方式还可以用到其他地方。    飞过矿区，可以寻找地表变化和矿化蚀变信息；飞过农田，可以观察作物长势和营养状况；换成其他模型，还可以寻找森林、水库和城市里的异常。    老颜也没料到，“天数天算”最早找到现实用途的地方之一，竟然是高光谱遥感。    原因并不复杂。    高光谱产生的数据足够多，星地通信又是天然的瓶颈。    卫星先把普通区域和无效信息筛掉，只传重点数据，有机会缩短响应时间，也能少占用一部分下行带宽。    太空算力到了这里，终于开始干具体的活了。         中国的星上智能力量正在形成          中科西光并不是国内唯一探索星上智能处理的团队。

J |     武汉一号同时具备高分辨率成像、高光谱成像和在轨智能图像处理能力。    经过在轨测试，这颗卫星于2024年9月正式开始提供数据服务。

K |     据《长江日报》报道，武汉一号还曾与福建一家遥感应用企业签订价值1800万元的数据服务采购订单。    此前发射的地卫智能应急一号，也开展过星载智能载荷和相关算法的在轨验证。    这些卫星走的路线并不完全一样，但至少说明，这条路不是中科西光一家在走。    西光贰号01星的特点，是把定量高光谱、248 TOPS星载算力和行业模型装进了同一颗商业卫星。    按照中科西光的表述，它是国内首颗“定量高光谱智算卫星”。

L |     这个名字确实有点长，却不能随便省略。    中国早已有高光谱卫星，也做过高光谱数据在轨处理。    少了“定量”“高光谱”或者“智算”中的任何一个词，意思都会发生变化。    西光贰号01星准备做的，是在获取高光谱数据以后，继续完成初步计算和识别，再把更值得关注的内容传回地面。         中科西光还要继续往下发          中科西光今年4月曾公开表示，已有11颗卫星组网运行。    随着西光贰号01星入轨，它的在轨队伍又增加了一颗。    按照公开规划，“西光系列”最终将由158颗卫星组成，包括108颗通用高光谱及红外监测卫星、40颗双碳监测高光谱卫星和10颗功能型卫星。

M |     西光贰号01星，是定量高光谱智算星座的首发星。    这也意味着，248 TOPS不是一次单独的技术展示。    后续卫星如果继续沿用高光谱载荷、星载计算和行业模型，中科西光想做的，就是把这种能力逐步铺到更多区域和更多任务中。    不过，一颗星成功入轨只是开头。

N |     真正麻烦的，是卫星换一条河、换一片矿区之后，模型还能不能认得准。    样本要重新补，参数也要跟着调，这些都得靠一次次实际任务磨出来。         客户真正需要的，是结果          高光谱卫星可以获取大量数据，但矿山、水利和农业部门通常不会自己研究一条条光谱曲线。    他们更关心的是：哪里出现了异常，变化到了什么程度，哪些地方需要尽快派人去看。    西光贰号01星先在轨道上筛选数据、标记重点区域，地面再做精细处理。    卫星先把重点区域圈出来，客户不用等所有数据都传完，就能先知道哪些地方值得去查。    云南省地质矿产勘查开发局参与了彩云高光谱01星建设，也给这颗卫星安排了明确的应用场景。    云南矿产资源丰富，河流、森林和高原农业的类型也很复杂。

o |     卫星入轨后，水体泥沙、矿区变化等模型将先在这些真实场景中接受检验。    以后能不能用到更多地区，要看模型换了环境以后是否依然准确，也要看星上筛选究竟能为实际工作节省多少时间。

p |     先把云南的任务干好，比提前讲多大的市场更重要。         这一步，值得鼓掌          西光贰号01星刚刚入轨，接下来还要完成平台测试、载荷开机和模型验证。    248 TOPS在轨运行是否稳定，行业模型能做到什么程度，星上处理究竟能省下多少时间和通信资源，都要看后续结果。    但这次发射至少把一件事做实了。    太空算力未必一开始就要训练大模型，也不必急着把整座数据中心搬上太空。    先让它看一条河流、一片矿区或者一块农田，把异常位置挑出来，再将重点数据传回地面，同样是一条现实的路。    遥感卫星拍完地球，通常还要等地面慢慢解读。

q |     西光贰号01星准备多做一步：先在天上看一看，再算一算。    对中国民营航天来说，这样的进步不一定轰轰烈烈，但很实在。    太空算力先有活干，比先讲多大的故事更重要。

Current article:http://v88zm.zongchaeiwengshuabaisisuofu.sbs/bxe7x9h/20260826/4124.html

Published on:00:15:33