AntVideoRecord: Autonomous system to capture the locomotor activity of leafcutter ants

Low-cost autonomous system produces less subjective results when recording leafcutter ants

Specifications

First model Material: ABS and PET-G
Second modelAnchor point: smooth metal rod (diameter: 10 mm)
Third modelMaterial: ABS plastic box
Size: 150x100x100 mm
Features: waterproof dustproof IP65
ComputerRaspberry Pi ZeroW (RPiW)
Raspberry Pi ZeroW (RPiW)size: 65x30 mm
Processor: 1 GHz BCM2835 single-core
RAM: 512 MB
Operating systemLinux
ConnectivityWi-Fi and Bluetooth
Storage16 GB microSDXC card
Camera module8 MegaPixel Raspberry Pi camera v2 (with a 110 lens); 1080p
Power12 V gel battery
total capacity: 25,000 mAh
regulator: 5V 3A Step-Up/Down
Energy recharging20W polycrystalline solar panel with a nominal charge of 1.24 Ah
Lighting systemPassive: 850 nm infrared led diodes
Active: white led lights with 5 mm diameter
Default configuration
Recording six 30-minute videos with the maximum capture area of the lens
Resolution640x480 pixels
Records format .mp4 format

Overview

This tech spec was submitted by Sabattini Julian Alberto as part of the University Technology Exposure Program.


Problem / Solution

Leaf-cutting ants (LCA) are an ecological keystone group in tropical ecosystems. However, some LCA species appear to take advantage of anthropogenic landscape alterations and become problematic pests. It is thus necessary to understand the responses of LCA species to anthropogenic environments and their response to human intervention and climatic change. People use automatic monitoring systems to analyze the videos in slow motion manually. Unfortunately, these solutions cannot be used continuously because of battery and storage limitations.

AntVideoRecord (AntVRecord) is an innovative, precise, and simple to manufacture and use tool that monitors the behaviors of LCA under varying field conditions. This will help understand the ant's field and lab behavior. It can also be used as an applied agronomic application technology as a decision criterion for sustainable IPM.

Design

Mechanical Design

The equipment's surroundings may be humid and hot. To protect the electronics, watertight support was needed. Three models were developed and evaluated on field conditions. The first model was based on rectangular support that was 3D-printed using ABS and PET-G. The top attachment points are metal rods. The indication for the user of LED lights and energy input are on one side. A retractable lid protects the video storage system. This model has trouble with rain and summer heat. The number of support points does not provide a perpendicular image to the forage trail since the support points are unchangeable.

A second design of the structure was 3D-printed. It features increased movement freedom and a 10 mm metal rod anchor point. This design allows landscape-based camera orientation. Its larger internal volume reduces electrical overheating in hot conditions. No moisture damage was present. However, the high humidity made the fulcrum less rigid, causing unwanted wind movements. Additionally, this design demanded smaller electronic plates, which means that tools may not be available to many users, requiring increasing manufacturing costs.

The final model was based on a waterproof, dustproof IP65 ABS plastic box. It's a simple shape that can be adapted from market products or printed. Only a few cover screws must be removed to change the storage device. Although this model is waterproof, it still is improved with a 3D-printed watertight cover. Two equipment-securing bolts were placed on the box and glued with sealant. The box's bottom was cut out to hold the camera module. The lighting arrangement also required eight 5-mm circles. The box is anchored to the ground through two points, offering stability when recording while permitting vertical movement to alter the recording area. The box's straightforward assembly simplifies component replacement.

 

Electronic Design

The core of the AntVRecord is a low-cost single-board computer with the addition of a Raspberry Pi Camera and other low-cost electronic components. RPiW has a 1 GHz single-core BCM2835 processor with 512 MB RAM, which may be configured as a desktop computer using an HDMI and USB port to an external device. Its embedded Wi-Fi and Bluetooth cards allow wireless and mobile monitoring of AntVRecord in the field.

All recorded data are stored in the MicroSDXC card. It can take HD video, and digital still images with its 8 MegaPixel Raspberry Pi camera module v2. A 20w polycrystalline solar panel gives the system autonomy that helps prevent spike damage. The lighting system uses a series of light sensors to turn on when natural light is scarce, which can be passive or active.

 

Design of the graphical user interface: AntProject

AntVRecord uses a graphical interface for RPiW's Linux OS. The OS is Raspbian Buster desktop distribution, which has minimal applications loaded, leaving space for video capture and storage. For video recording and storage, a 2.0 pen drive connected to the RPiW's Mini-USB port is used. AntVRecord OS is installed manually on a Mini-SD card and then using a graphical interface.

AntProject was developed to configure the recording and storage system. It can be accessed remotely using the VNC Viewer plug-in through Bluetooth-paired phones, tablets, or laptops. The communication system requires Blueman2, a GTK Bluetooth Manager that simplifies operations like controlling the BlueZ API and connecting to Bluetooth networks and pairing. It's Python-based, lightweight, and GPL licensed. The VNC Viewer server can be used remotely to edit AntProject configuration from a cabinet without being on-site at the equipment installation site.

The devices were set to record video at 30 frames per second at two resolutions: 640x480 and 1200x720. 26Ah battery lasted 26 days in bright sunlight, clouds, and rain. It exceeded energy stability standards. No moisture or water infiltration inside the watertight box could damage the electronics. The minimum and highest video recording durations per deployment were 15 h (PNA La Esmeralda) and 854 h (School of Agronomy, UNER).

References

A research paper describing the challenge, design, and outcome of the research.

J.A. Sabattini, J.M. Reta, L.A. Bugnon, J.I. Cerrudo, R.A. Sabattini, A. Peñalva, M. Bollazzi, M.O. Paz, F. Sturniolo

Wevolver 2023