The client is looking for someone to develop an API that will accept 2-3 or more if needed to accurately calculate real images of a certain object in a warehouse environment from different angles and return its complete dimensions.
The object should be placed inside a designated area on the floor which will be included in the image. Only one object will be placed inside the designated area. The client will provide the real dimensions of the designated area as marked on the floor.
The object itself will be a pallet that includes boxes. The boxes can overlap the pallet as therefore the client would need to calculate the maximum length, width and height of the pallet. Accuracy should be at 3/4" tolerance (around 2cm).
Attached sample images you can work with. The final API will need to accept real images taken from within the warehouse. The client will need to detect the square of the designated area, detect the pallet inside and measure its length, width and height.
We have done quite a few image processing web/mobile apps. We have used custom algorithms to derive the area, slope, volume based on the input data from the image.
Here is the latest project, which calculates the area of the roof based on its slope and perspective size compared to the height of the door.
Another project involves detecting the presence of employees at their work stations in order to calculate their working hours. The application was trained using around 500 positive and negative image samples and after that, it was able to detect the presence/absence of an employee at his workstation.
The third one was identifying the wood logs from the image uploaded. We needed to get the precise measurement of objects' size. The target was to identify sections of wood logs, mark circles on each identified log, in turn. Next was to get the width, height, and length of the wood logs and calculate the weight.