Most current research primarily focuses on improving experimental accuracy using large models, often neglecting the deployment challenges. There is a growing need for lightweight algorithms in certain remote sensing devices. Moreover, remote sensing images (RSIs) often contain numerous small, densely distributed targets, which pose significant detection challenges. To address these issues, we have improved the YOLOv8s network and developed a lightweight remote sensing object detection (RSOD) network based on multiscale features fusion and context information (MFFCI-YOLOv8). This network combines multiscale feature fusion and contextual information to accurately detect objects in RSIs. Firstly, we introduce the lightweight CSP bottleneck with attention (LCA) module, which utilizes partial convolution calculation and SimAM attention mechanisms to decrease the number of param-eters and computational complexity while enhancing feature extraction capabilities. Secondly, we design the gate spatial pyramid pooling fast (GSPPF) module to enhance the model's perception of scale and contextual information, thus improving the detection of small objects. Lastly, we employ the multiscale fusion lightweight neck (MFLNeck) module for more efficient multiscale feature fusion, preventing the loss of small objects. Compared to YOLOv8s, our overall model reduces the number of parameters by 7.7% and FLOPs by 11.9%. We validated the accuracy of MFFCI-YOLOv8 on two remote sensing datasets, NWPU VHR-10 and VisDrone. The experimental results demons-trate that our model offers a low computational cost and high detection accuracy compared to other RSOD models and other YOLO models.

As a commonly used food additive, sulfite (SO₃²⁻) is popular with food manufacturers due to the functions of bleaching, sterilizing, and oxidation resistance. However, excess sulfites can pose a threat to human health. Therefore, it is particularly important to achieve rapid and sensitive detection of SO₃²⁻. Herein, a colorimetric sensor was invented for visual, meticulous, and rapid detection of SO₃²⁻ based on MIL-53(Fe/Mn). Bimetallic nanozyme MIL-53(Fe/Mn) was prepared by a one-pot hydrothermal reaction. The prepared MIL-53(Fe/Mn) can effectively catalyze the oxidation of colorless TMB to a blue oxidation product (oxTMB). The introduction of SO₃²⁻ causes significant discoloration of the reaction system, gradually transitioning from a visible blue color to colorless. Hence, a sensitive colorimetric sensor for SO₃²⁻ detection was developed based on the decolorization degree of the detection system. Further, the discoloration was ascribed to the inactivation of nanozyme and the strong reducing ability of SO₃²⁻. Under the optimal experimental conditions, there was a good linear relationship between the absorbance at 652 nm and SO₃²⁻ concentration in the linear range of 0.5–6 μg mL⁻¹ with a limit of detection (LOD) of 0.05 μg mL⁻¹. The developed method was successfully applied to the detection of actual samples of white wine with good accuracy and recovery. Compared to traditional methods, this colorimetric sensor produces similar detection results but significantly reduces the detection time. Compared to traditional methods, this colorimetric sensor can not only reduce the detection costs effectively but also help the food industry maintain quality standards. Strong anti-interference capability, simple operation, and low detection limits ensure the excellent performance of the colorimetric sensor in detecting SO₃²⁻ in white wine. The combination of a smartphone and a colorimetric analysis application has also greatly facilitated the semi-quantitative, visual on-site detection of SO₃²⁻, which has opened up an application prospect of an MIL-53(Fe/Mn)-based detection platform. Our work has indicated a new direction for the detection of SO₃²⁻ and provided important assurance for food safety.

To address the critical challenges in Synthetic Aperture Radar (SAR) ship target detection, including complex background speckle noise interference and the difficulty in balancing model lightweight design with detection accuracy, this paper proposes an innovative PPDM-YOLO model. Through modular architecture design, we establish a four-part technical framework: First, a lightweight feature extraction module named PCA is developed to reduce computational complexity by analyzing feature map redundancy, effectively mitigating feature degradation caused by noise. Second, the Noise-Resistant Enhancement Module, PSA-G, integrates the Multi-Scale Adaptive Gradient Threshold (AGT) module with a Dynamic Spatial Attention mechanism. This integration enhances target feature representation while effectively suppressing noise interference. Third, DySample technology is employed in place of conventional upsampling methods to improve the quality of feature reconstruction and preserve spatial details. Additionally, a Multi-Scale Fusion Small Target detection network (MSTFNet) is introduced to boost small object detection through cross-layer feature interaction. Experimental results on HRSID and SSDD datasets demonstrate that PPDM-YOLO achieves 93.7% mAP50 and 70.3% mAP50-95 on HRSID, while reaching 99.4% mAP50 and 78.7% mAP50-95 on SSDD, showing significant advantages over mainstream detection models.With 34.7% fewer parameters than YOLOv11n, our model achieves optimal balance among noise suppression, model lightweighting, and detection accuracy. This research provides an efficient and reliable technical solution for real-time SAR ship detection in complex marine environments.

Detecting small and occluded objects in complex traffic scenes remains a major challenge for intelligent transportation systems, including the difficulty in balancing high accuracy and low computational cost. To address these challenges, this paper proposes an innovative MSF-YOLOv11n model. With a modular architecture design, we propose three innovative technical frameworks. First, the Adaptive Attention-Enhanced Wavelet Module (AAEWM) performs multi-scale wavelet decomposition combined with attention-guided reconstruction, enabling the model to capture structural information such as edges and textures while suppressing noise interference. Second, the Feature-Enhanced Attention Upsampling Module (FEAUM) integrates lightweight Ghost convolution with Shuffle Attention and employs a hybrid interpolation and convolution strategy, which preserves fine-grained details in the upsampling process. Third, the Multi-Scale Fusion Lightweight Triple Attention Module (MFTAM) employs depthwise separable convolutions and a triple attention mechanism across channel, spatial and positional dimensions, effectively enhancing multi-scale feature interaction and improving small-object localization accuracy. Compared with YOLOv11n, MSF-YOLOv11n improves mAP50-95 by 3.8% and mAP50 by 3.7% on UA-DETRAC, and gains 3.6% in mAP50-95 and 3.9% in mAP50 on DAIR-V2X, all with a computational cost of only 6.5 GFLOPs and an inference speed of 100 FPS on UA-DETRAC. These results demonstrate that MSF-YOLOv11n achieves a strong balance of accuracy and efficiency for real-world traffic perception.