▪ Development of algorithm to detect the defects on both the larger and smaller surface of the Yoke ▪ Algorithm shall be suitable to detect the defects on the surface as well as the edges of the surface ▪ Algorithm shall be suitable to identify each specific defects. If no defect is identified then the part can be classified as a OK part. ▪ Testing and implementation of algorithm on live production and ensuring near to 100% accuracy of inspection as compared to the existing manual inspection methodology ▪ The algorithm shall be capable of inspecting parts at the rate of ~10 parts per minute. ▪ The samples will be stationary while the images is being captured.
▪ The algorithm shall be later developed into an application software which is easy to handle for a production scenario. ▪ The software shall have relevant interface to interact and maintain. ▪ The software developed shall communicate with the vision system to capture the image and be able to read the image. ▪ The software shall have a feature to communicate and enable connection with an industrial PLC.
▪ The algorithm shall be triggered by an external input (24VDC) and shall provide multiple output based on the identified error. ▪ The error name definition shall be dynamic and configurable.
- Needs discussion before consideration
▪ The algorithm shall be developed in an authorized software or platform which can later be transferred to L&T and shall comply with all statuary requirements. ▪ The hardware for capturing the images and processing the above developed algorithm shall be in scope of L&T. ▪ The hardware specifications suitable for the computational requirements of the developed algorithm shall be used by L&T. The required hardware specifications shall be conveyed after the successful trials of the algorithm. ▪ L&T shall provide the raw images required as the input for the development and implementation of the vision inspection algorithm. ▪ Other points as mentioned further in the document shall be part of the deliverable.
Design a vision-based solution towards automated inspection of PMR yoke. The objective of the project is: ▪ To develop a vision-based solution towards inspection of Yoke, and towards this, o Develop an algorithm to classify the yoke into accepted (satisfying the benchmark) and rejected samples based on captured visual information. o Develop an algorithm to identify the type of defect in the rejected yoke sample (i.e, samples with scratches, dents, burr, white/black spots, grinding lines, and uneven ground surfaces) and label the rejected yokes as workable and non-workable rejected samples based on type of defect. ▪ To develop a interface for communication with the physical system, and towards this, o Develop an application software which is easy to handle for a production scenario. o Develop an interface that can communicate and enable connection with an industrial PLC.
