Due to large project size of hydropower engineering, construction quality inspection is normally conducted by overseeing, with supplementary monitoring and data evaluation when available. With the rapid advancement of artificial intelligence in recent years, intelligent technology has been increasingly applied in the practice of engineering construction and inspection. To improve the efficiency and intelligence of construction quality inspection, an intelligent bi-directional verification system based on algorithms, such as ZoeDepth and EDTER, was developed. First, a virtual field of the physical world is generated using the depth estimation approach. Image processing techniques, such as image object recognition and object delineation identification, are used to establish the location, position and posture of actual objects in images as well as measure their dimensions. Second, object-to-object matching is accomplished by aligning the engineering design in BIM models with the objects identified in the images. As a result, image processing models can be calibrated for intelligent object recognition. Object dimensions can then be measured in the virtual field. Corresponding object design information can be retrieved from the BIM model for construction quality examination. Finally, an inspection report is automatically generated and can be synchronized to a comprehensive engineering management platform, when network communication is available. Based on the aforementioned technical framework, a single-person intelligent inspection wearable was developed and tested for technical evaluation. The results indicated that the proposed algorithms can quickly identify construction objects for in-situ comparison with the design in BIM models and the device satisfies the requirement of being simple to wear and operated by one person when perform the engineering inspection.