Research on Weld Defect Recognition Method Based on Mask R-CNN

Enlai Yang , Zirui Zhao — Mon, 15 Jun 2026 00:00:00 +0000

Automatic detection of weld defects is of great significance to ensuring the quality of industrial products and engineering safety. To address the challenges of low contrast, large defect scale variation, complex background noise, and limited sample data in ultrasonic weld defect images, an improved Mask R-CNN instance segmentation model is proposed. First, the original ResNet+FPN backbone network is replaced with the RSU7 multi-scale feature extraction module to enhance the model’s ability to capture details of tiny defects through a nested U-structure and residual connections. Second, the CBAM attention mechanism is connected in series between the backbone network and the region proposal network to suppress background noise and highlight defect regions in both channel and spatial dimensions. Experiments are conducted on an ultrasonic weld defect dataset containing only 105 images. The results show that the improved model achieves a mean average precision (mAP) of 0.7564, which is 4.2% higher than that of the baseline Mask R-CNN (ResNet50+FPN). The coefficient of variation (CV) is 2.26%, indicating better training stability than the comparison models. The maximum AP drop rate under image disturbance is only 9.1%, demonstrating strong robustness. The proposed method realizes high-precision and high-stability defect detection and segmentation in small-sample scenarios, providing an effective technical solution for industrial ultrasonic welding quality inspection.

Effects of Multiple Cropping with Green Manure on Yield and Agronomic Traits of Spring Wheat

Ranran Guo — Mon, 15 Jun 2026 00:00:00 +0000

A split-plot experiment was conducted to investigate the effects of green manure returning methods on wheat yield and phosphorus (P) accumulation. The main plots consisted of two fertilization levels for the subsequent wheat crop: no chemical fertilizer (N0) and chemical fertilizer application (N1, N 157.5 kg/ha + P₂O₅ 78.75 kg/ha). The sub-plots comprised three green manure returning methods: no green manure (G0), green manure with root stubble retention (G1), and green manure with full incorporation (G2). Wheat yield, agronomic traits, and P content in various wheat organs were measured after harvest. The results showed that green manure returning significantly increased the yield and P accumulation of the subsequent wheat crop. Compared with chemical fertilizer alone (N1G0), the combination of green manure and chemical fertilizer significantly improved wheat yield. Specifically, the full incorporation treatment (N1G2) and root stubble retention treatment (N1G1) increased yield by 8.7% and 5.8% in 2024, and by 7.2% and 5.2% in 2025, respectively. The combined application of green manure and chemical fertilizer synergistically optimized agronomic traits and yield components, including grains per spike, thousand-grain weight, spike number, stem diameter, and plant height. Path analysis indicated that thousand-grain weight was the dominant factor affecting yield. Therefore, green manure full incorporation combined with chemical fertilizer is the optimal green manure-chemical fertilizer pattern for improving phosphorus availability in Qinghai region.

Journal of Research in Multidisciplinary Methods and Applications

Research on Weld Defect Recognition Method Based on Mask R-CNN

Effects of Multiple Cropping with Green Manure on Yield and Agronomic Traits of Spring Wheat