{"id":281,"date":"2026-01-11T10:38:08","date_gmt":"2026-01-11T02:38:08","guid":{"rendered":"https:\/\/blog.zeronely.cn\/?p=281"},"modified":"2026-01-11T10:38:10","modified_gmt":"2026-01-11T02:38:10","slug":"%e5%85%b3%e4%ba%8ers%e5%a4%a7%e5%9b%be%e5%a4%84%e7%90%86%e7%9a%84%e8%ae%ba%e6%96%87%e8%b0%83%e7%a0%94","status":"publish","type":"post","link":"https:\/\/blog.zeronely.cn\/index.php\/2026\/01\/11\/%e5%85%b3%e4%ba%8ers%e5%a4%a7%e5%9b%be%e5%a4%84%e7%90%86%e7%9a%84%e8%ae%ba%e6%96%87%e8%b0%83%e7%a0%94\/","title":{"rendered":"\u5173\u4e8ers\u5927\u56fe\u5904\u7406\u7684\u8bba\u6587\u8c03\u7814"},"content":{"rendered":"\n

GeoLLaVA-8K: Scaling Remote-Sensing Multimodal Large Language Models to 8K Resolution<\/h3>\n\n\n\n

\u63d0\u51faBackground Token Pruning\u7b56\u7565\uff0c\u901a\u8fc7\u8bed\u4e49\u76f8\u4f3c\u6027\u805a\u7c7b\u6709\u6548\u538b\u7f29\u5197\u4f59\u80cc\u666ftokens\u3002\u8bbe\u8ba1Anchored Token Selection\u65b9\u6cd5\uff0c\u5229\u7528[CLS] token\u6ce8\u610f\u529b\u56fe\u4fdd\u7559\u5173\u952e\u76ee\u6807\u4fe1\u606ftokens\u3002<\/p>\n\n\n\n

Co-Training Vision Language Models for Remote Sensing Multi-task Learning<\/h3>\n\n\n\n

\u9996\u4e2a\u540c\u65f6\u652f\u6301\u5e38\u89c4\u548cUHR\u9065\u611f\u56fe\u50cf\u7684\u7edf\u4e00VLM\u3002\u6db5\u76d6\u5206\u7c7b\u3001\u68c0\u6d4b\u3001\u5b9a\u4f4d\u3001\u95ee\u7b54\u7b49\u591a\u4efb\u52a1\u3002\u7a81\u7834\u73b0\u6709\u6a21\u578b\u4ec5\u4e13\u6ce8\u5355\u4e00\u5206\u8fa8\u7387\u57df\u6216\u4efb\u52a1\u7c7b\u578b\u7684\u9650\u5236\uff0c\u5b9e\u73b0\u771f\u6b63\u7684\u591a\u4efb\u52a1\u5b66\u4e60\u3002<\/p>\n\n\n\n

XLRS-Bench: Could Your Multimodal LLMs Understand Extremely Large Ultra-High-Resolution Remote Sensing Imagery?<\/h3>\n\n\n\n

\u4e00\u4e2a\u7efc\u5408\u57fa\u51c6\u6d4b\u8bd5\uff0c\u7528\u4e8e\u8bc4\u4f30MLM\u5728\u8d85\u9ad8\u5206\u8fa8\u7387RS\u573a\u666f\u4e0b\u7684\u611f\u77e5\u4e0e\u63a8\u7406\u80fd\u529b\u3002
\u76ee\u524d\u6700\u5927\u3001\u6700\u6e05\u6670\u7684\u9065\u611f\u591a\u6a21\u6001 benchmark \u6d89\u53caVQA\uff0cgrounding<\/p>\n\n\n\n

When Large Vision-Language Model Meets Large Remote Sensing Imagery: Coarse-to-Fine Text-Guided Token Pruning<\/h3>\n\n\n\n

\u805a\u7126\u4e8e\u5927\u5c3a\u5bf8\u9065\u611f\u56fe\u50cf\u7684\u5197\u4f59token\uff0c\u63d0\u51fa\u4e00\u79cd\u7c97\u5230\u7ec6\uff08Coarse-to-Fine\uff09\u6587\u672c\u5f15\u5bfc\u7684 Token \u526a\u679d<\/strong>\u6846\u67b6\uff0c\u901a\u8fc7\u6587\u672c\u67e5\u8be2\u7684\u8bed\u4e49\u4fe1\u606f\u5f15\u5bfc\uff0c\u5206\u9636\u6bb5\u7b5b\u9009\u9065\u611f\u5f71\u50cf\u4e2d\u7684\u6709\u6548\u89c6\u89c9 token\uff0c\u5254\u9664\u5197\u4f59\u80cc\u666f token
\u6784\u5efa\u4e86\u4e00\u4e2abenchmark LRS-VQA<\/p>\n\n\n\n

GEOBench-VLM: Benchmarking Vision-Language Models for Geospatial Tasks<\/h3>\n\n\n\n

\u9762\u5411\u5730\u7406\u7a7a\u95f4\u4efb\u52a1\u7684\u591a\u6a21\u6001\u8bc4\u4f30\u57fa\u51c6 GEOBench-VLM\uff0c\u8bbe\u8ba1\u4e86 6 \u7c7b\u6838\u5fc3\u5730\u7406\u7a7a\u95f4\u4efb\u52a1\uff08\u5305\u62ec\u5730\u7269\u5206\u7c7b\u3001\u76ee\u6807\u68c0\u6d4b\u3001\u89c6\u89c9\u63a5\u5730\u3001\u5730\u7406\u7a7a\u95f4 VQA\u3001\u8de8\u6a21\u6001\u68c0\u7d22\u3001\u591a\u8f6e\u5bf9\u8bdd\u63a8\u7406\uff09<\/p>\n\n\n\n

GeoChat: Grounded Large Vision-Language Model for Remote Sensing<\/h3>\n\n\n\n

\u63d0\u51fa GeoChat\uff0c\u9996\u4e2a\u652f\u6301\u9ad8\u5206\u8fa8\u7387\u9065\u611f\u56fe\u50cf\u591a\u4efb\u52a1\u5bf9\u8bdd\u7684 VLM\uff0c\u53ef\u5904\u7406\u56fe\u50cf\u7ea7\u4e0e\u533a\u57df\u7ea7\u67e5\u8be2\uff0c\u5e76\u901a\u8fc7\u7a7a\u95f4\u5750\u6807\u5b9e\u73b0\u76ee\u6807\u89c6\u89c9\u5b9a\u4f4d\uff0c\u6784\u5efa\u9065\u611f\u591a\u6a21\u6001\u6307\u4ee4\u6570\u636e\u96c6\uff0c\u57fa\u4e8e LLaVA-1.5 \u5fae\u8c03\uff0c\u5efa\u7acb\u591a\u4efb\u52a1\u5bf9\u8bdd\u57fa\u51c6\u5e76\u9a8c\u8bc1\u96f6\u6837\u672c\u6027\u80fd\u3002<\/p>\n\n\n\n

GeoGround: A Unified Large Vision-Language Model. for Remote Sensing Visual Grounding<\/h3>\n\n\n\n

\u9996\u4e2a\u7edf\u4e00\u652f\u6301\u9065\u611f\u89c6\u89c9\u5b9a\u4f4d\u4e2d\u6c34\u5e73\u6846\uff08HBB\uff09\u3001\u5b9a\u5411\u6846\uff08OBB\uff09\u4e0e\u5206\u5272\u63a9\u7801\uff08mask\uff09\u4e09\u7c7b\u8f93\u51fa\u7684\u591a\u6a21\u6001\u5927\u6a21\u578b\uff0c\u901a\u8fc7 Text-Mask \u8303\u5f0f\u5c06\u50cf\u7d20\u7ea7\u4fe1\u606f\u8f6c\u5316\u4e3a\u6587\u672c\u5e8f\u5217\uff0c\u914d\u5408\u6df7\u5408\u76d1\u7763\u8bad\u7ec3\u4e0e\u5927\u89c4\u6a21\u6307\u4ee4\u6570\u636e\u96c6 refGeo\uff0c\u5b9e\u73b0\u7075\u6d3b\u8f93\u51fa\u4e0e\u8de8\u4efb\u52a1\u4e00\u81f4<\/p>\n\n\n\n

TinyRS – R1: Compact Multimodal Language Model for Remote Sensing<\/h3>\n\n\n\n

\u9762\u5411\u9065\u611f\u8f7b\u91cf\u5316 MLLM \u9700\u6c42\uff0c\u63d0\u51fa TinyRS – R1\uff0c\u901a\u8fc7\u6a21\u578b\u84b8\u998f\u3001token \u7a00\u758f\u5316\u4e0e LoRA \u5fae\u8c03\uff0c\u5728\u4fdd\u8bc1\u6027\u80fd\u7684\u524d\u63d0\u4e0b\u538b\u7f29\u6a21\u578b\u89c4\u6a21\u4e0e token \u6570\u91cf\uff0c\u652f\u6301\u7a7a\u95f4\u5b9a\u4f4d\u4e0e\u573a\u666f\u7406\u89e3\u7b49\u4efb\u52a1<\/p>\n\n\n\n

VisionZip: Efficient Token Compression for Vision – Language Models<\/h3>\n\n\n\n

\u9488\u5bf9\u89c6\u89c9 token \u5197\u4f59\u5bfc\u81f4\u7684\u8ba1\u7b97\u6210\u672c\u9ad8\u95ee\u9898\uff0c\u63d0\u51fa VisionZip\uff0c\u901a\u8fc7\u4e0a\u4e0b\u6587 token \u5408\u5e76\u4e0e\u4fe1\u606f\u6027 token \u9009\u62e9\uff0c\u51cf\u5c11\u89c6\u89c9 token \u6570\u91cf\uff0c\u5728\u4fdd\u6301\u6027\u80fd\u7684\u540c\u65f6\u63d0\u5347\u63a8\u7406\u901f\u5ea6\uff0c\u9002\u914d\u56fe\u50cf\u4e0e\u89c6\u9891\u7406\u89e3\u4efb\u52a1<\/p>\n","protected":false},"excerpt":{"rendered":"

GeoLLaVA-8K: Scaling Remote-Sensing Multimodal Large La […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-281","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/posts\/281","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/comments?post=281"}],"version-history":[{"count":1,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/posts\/281\/revisions"}],"predecessor-version":[{"id":282,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/posts\/281\/revisions\/282"}],"wp:attachment":[{"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/media?parent=281"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/categories?post=281"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.zeronely.cn\/index.php\/wp-json\/wp\/v2\/tags?post=281"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}