Research On AR Guidance-Based Visual Inspection Method For Cable Laying Quality Consistency Check

Aerospace cables play a role in transmitting signals and power for aerospace products.Thus,cable laying quality directly exerts great influence on products properties.At present,cable laying mainly depends on manual operation because of aerospace products' complex structure.Aerospace products' compact configuration also makes cable laying space or cable assembly space narrow.It causes inconvenience in controlling cable laying orientation.What's more,workers' technical ability will affect cable bending extent in laying process.These factors make it difficult to ensure cable laying quality consistency.As a consequence,cable laying process needs to be monitored and analyzed,making cable assembly quality meet requirements by AR aided assembly technology.In view of the above problems and research objectives,studies were carried on as follows: cable target image segmentation,quality feature extraction for cable in laying process,product information modeling for cable and system development.Proposes cable target image segmentation method for manual cable laying process.Firstly,cable was detected in a given cable image from egocentric vision and all minimum rectangular bounding boxes for cable target regions were calculated.Meanwhile,thumb points were detected according to workers' hand area extracted from cable image.Then,the positional relations between thumb points and all rectangular bounding boxes were estimated on the basis of 2D orientation bounding box intersection principal.In this way,the target bounding box belonging to cable target region could be found.Finally,we can segment cable target image in accordance with the target bounding box.Proposes a method for cable laying feature extraction and match.Cable target image was thinned to obtain its skeleton information with burr removal algorithm.These skeleton information were transformed into 3D coordinates with depth information of cable target image.The 3D coordinates of skeleton were used to compute bending curvature and bending radius for skeleton.Meanwhile,feature of cable image from egocentric vision was extracted to match with model cable image and the number of matched points which belong to cable target area in model cable image was computed.The number was utilized to measure the similarity of cable bending orientation.Proposes a product information model for cable laying process.Cable information model,including part information,cable laying process information,cable laying quality detection information,simulated video information of cable laying process and et al,were established and connections among models were built to provide related data in AR assistance system for cable assembly.Based on above research,CAD models of cable and guidance information were established and an intelligence guidance system for cable laying was developed.The system was proved effective with electronic module simulated cable assembly case.