Challenge

The task is to validate the technical specifications for proper installation of e-batteries during assembling process of electric cars. Some of the specifications can be formulated as the following questions: a) are the plugs properly closed? b) are the cables straight or twisted? c) do the cables lie in the right spot? c) are some parts missing? etc.

Solution

We trained a group of deep neural networks for detecting the keypoints of interest like plugs, locks and cables and to validate their corresponding state.

Impact

The proposed approach was selected to be applied for automatic validation of multiple steps during the production process.