Engineering student created an impressive DIY cable cam with AI-powered tracking: Digital Photography Review

The Cablecam is constructed round a single brushless motor and Raspberry Pi four with an Arducam connected to a two-axis gimbal.

Uruguayan engineering pupil, Maximiliano Palay, has shared an in depth breakdown exhibiting how he constructed Cablecam, a DIY cable digicam. Whereas this isn’t precisely your weekend Raspberry Pi mission, when you’ve got the engineering chops (in addition to money and time), it’s one thing you may piece collectively your self.

Max created the design for the Cablecam utilizing Autodesk Fusion 360 (through an Schooling License).

The Cablecam, as Max calls it, was the primary mission of his to combine all of his disciplines collectively: ‘software program, design and a few fabrication strategies’ and was his first try at making a shifting robotic. Max says he made a lot of the Cablecam from elements he had ‘laying round,’ however except you’ve got entry to a 3D printer and a CNC machine, and have a number of pocket computer systems and motors sitting round, you would possibly have to buy a number of gadgets. Under is a listing of the important thing parts Max used to create the Cablecam:

Like different cable (or line) cameras, the Cablecam can run alongside a string or cable in a single axis to offer a dolly-style impact with out the necessity to arrange a sophisticated rail system. The body of the Cablecam is constructed of CNC-milled wooden items with 3D-printed gears customized made to maximise torque from the brushless motor. Under is a video exhibiting the gears in motion with out the motor connected:

To drive the motor, Max used an digital pace controller he had sitting round, but it surely was leftover from a drone mission, so it didn’t supply a reverse choice. To repair this, Max added a polarity change utilizing 4 relays with an Arduino in order that when the polarity is switched, the Cablecam strikes in the wrong way. The batteries used to energy the Cablecam have been ‘harvested cells’ taken from an outdated laptop computer.

The digicam onboard the Cablecam is an Arducam inside a customized 3D-printed case connected to a Tarot two-axis gimbal. Max used a wide-angle smartphone digicam lens adapter from a package in entrance of the Arducam to get a large angle-of-view. The complete Cablecam is powered by a Raspberry Pi four, which communicates wirelessly with the bottom unit.

The bottom station, which wirelessly controls the Cablecam, makes use of Nvidia’s Jetson Nano as the primary pc and is all packaged inside a Seahorse SE120 case. Along with the primary pc and fan used to maintain issues cool, the case additionally homes a display and potentiometer that’s used to manually management the Cablecam’s motion.

To take it a step additional, Max additionally built-in Nvidia’s pre-trained neural networks so as to add an computerized mode that identifies and tracks folks based mostly on sure standards. This takes footage despatched to the bottom unit from the Cablecam, analyzes it, then mechanically strikes the Cablecam to middle the topic within the body, as seen within the beneath demo video (Max apologized for the poor video high quality):

This isn’t your common DIY mission, but it surely goes to point out that even advanced items of digicam tools might be made when you’ve got the information and tools obtainable. If you happen to’re courageous sufficient to aim your individual Cablecam, you’ll want to shoot us a message so we are able to characteristic it as properly.

The total mission breakdown might be discovered over on Hackaday, the place Max shares a number of the sources he used to get every little thing up and working. Yow will discover extra of Max’s tasks on Hackaday and sustain together with his engineering feats through his LinkedIn profile.


Picture credit: Images by Maximiliano Palay, shared with permission.

Leave a Reply

Your email address will not be published. Required fields are marked *