Development and evaluation of a VR-based simulation platform for wild mushroom identification and risk perception

While Yunnan Province is hailed as China’s “Kingdom of Wild Mushrooms” for its abundant wild mushroom resources, the risk of poisoning remains a severe public health challenge. Frequent poisoning by accidental ingestion of toxic wild mushrooms poses a grave threat to public health and social stability. Traditional safety education materials, such as posters and manuals, are often found lacking because two-dimensional images cannot capture the three-dimensional structure and spatial relationships of key identification features like annuli and volvas, resulting in a superficial understanding. These methods of identification lack interactivity and situational immersion, fail to simulate decision-making in real foraging scenarios, and their safety warnings remain abstract, scarcely triggering users’ embodied risk perception. Crucially, the absence of real-time, quantitative assessment tools for tracking cognitive processes and identification logic leads to inaccurate and subjective evaluation of training efficacy. To address these limitations, this study develops a VR-based simulation platform for wild mushroom identification and risk perception, using innovative immersive technology to strengthen public identification skills and safety awareness.

Built on a layered, modular architecture comprising hardware, data, software, application, and user layers, the platform adheres to the software engineering principle of “high cohesion and low coupling” to ensure stability and scalability. Key technologies include four core components: high-fidelity virtual scene construction using a hybrid approach of real-scene scanning and manual refinement, combined with physically based rendering and level of detail technology to balance image quality and performance, with ecological layout of forest scenes and functional design of kitchen scenes based on real data; a task-guided intelligent interactive VR system where users can grab, rotate, and scale mushroom models, complemented by an intelligent prompt mechanism triggered by either 2.5 seconds of focused attention on key features or spatial proximity of similar species for comparison embodied risk warning that uses audio–visual effects to simulate the consequences of poisoning and strengthen safety memory; and a quantitative evaluation model built on full-process operation data, incorporating identification accuracy, efficiency, observation detail, and logicality to realize process-oriented assessment.

A controlled experiment was conducted with 60 participants with no background in botany or mycology. The subjects were randomly assigned to either an experimental group (n₁ = 30), which received at least 2 hours of VR platform training over one week, or a control group (n₂ = 30), which used traditional online learning methods. Post-test results showed significant improvements in both groups, but the experimental group achieved much greater gains—33.4 ± 9.1 points in theoretical tests and 39.5 ± 11.3 points in image recognition—than the control group’s 17.4 ± 8.7 and 16.3 ± 9.5 points, respectively. A strong positive correlation (r = 0.72) was observed between the experimental group’s identification accuracy and the duration of their observation of key features. Subjective feedback indicated high satisfaction, with immersion (4.8/5.0) and risk warning perception (4.7/5.0) receiving the highest ratings, confirming the platform’s effectiveness in improving memory and safety awareness.

This VR-based platform successfully integrates immersive environments, intelligent interaction, and process-oriented evaluation, effectively mitigating the shortcomings of traditional wild mushroom safety education. The platform significantly improves identification accuracy, safety awareness, and observation skills by facilitating situational cognitive experiences, active exploration opportunities, and emotional risk warnings. Its technical framework is highly transferable to other high-risk training fields such as food and drug safety, medical device identification, and industrial safety, featuring broad promotional value. Future developments will expand sample diversity, conduct long-term tracking of training effects, and optimize the platform with AI algorithms and haptic feedback devices, contributing to the overall improvement of social safety training efficiency.