For　the　problems　of　unsatisfactory　detection　accuracy　and　weak　real-time　performance　in　the　complicated　illumination　scenes　in　the　existing　deep　learning　target　detection　algorithmsan　anti-illumination　target　detection　network　model　YOLO-RLG　based　on　the　YOLO　algorithm　is　proposed.　Firstlythe　RGB　data　of　the　input　model　is　converted　into　HSV　dataand　the　S　channel　with　powerful　anti-illumination　capability　is　separated　from　the　HSV　data　and　fused　with　the　RGB　data　to　generate　RGBS　data　so　that　the　input　data　has　anti-illumination　capability.　Secondlythe　backbone　network　of　YOLOV4　is　replaced　with　Ghostnet　networkwith　the　model　assignment　ratio　between　ordinary　convolution　and　cheap　convolution　modified　to　improve　the　detection　efficiency　while　ensuring　the　detection　accuracy.　Finallythe　loss　function　of　the　model　is　improved　by　replacing　CIoU　with　EIoUwhich　enhances　the　target　detection　accuracy　and　algorithm　robustness.　The　experimental　results　based　on　KITTI　and　VOC　datasets　indicate　thatcompared　with　the　original　network　modelthe　FPS　improves　by　22.54　and　17.84　f/swith　the　model　reduced　by　210.3　Mthe　accuracyAPimproved　by　0.83%　and　1.31%and　the　algorithm′s　anti-illumination　performance　significantly　enhanced.　©　2023　SAE-China.　All　rights　reserved.