Skip to content

nikhil-x-gupta/hummingbird_detect

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

21 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

hummingbird_detect

This project's goal is to identify hummingbirds when they show up on a camera feed and notify the user/client when one is currently within the camera's view.

Screenshot

Table of Contents

Jetson Setup

Initially this was going to be set up on a Jetson Nano, initially running Jetpack 4.3.6. I have since migrated to using a Jetson Xavier running Jetpack 5.0.2. The following instructions and commands are helpful for updating Jetpack. In order to update all parts of the Jetpack SDK to the latest supported on Nano, which is 4.6.1. This version includes Linux version 32.7.3.

To check jetpack version type:

sudo apt-cache show nvidia-jetpack

To check Linux version:

dpkg-query --show nvidia-l4t-core

To update jetpack, follow the following steps:

First, edit etc/apt/sources.list.d/nvidia-l4t-apt-source.list to point to the desired repo using your preferred editor (just change the version to r35.1 in both lines) Then run:

sudo apt update
sudo apt dist-upgrade

Project Iterations

The first phase of the project is to set up a hummingbird classifier trained on existing dataset images of hummingbirds, other birds, gardens, and squirrels, to cover common features of the eventual deployment location (a backyard).

The second phase of the project is to integrate the classifier with a livestream of a separate camera with an RTSP feed.

The third phase of the project is to gather custom data for object detection and train an object detection model, working with the live RTSP feed.

The fourth phase of the project is to create an alert system or GUI to interact with the output of the detection.

I used VNC to facilitate working remotely on the Jetson Xavier via SSH and VLC Media Player to view the feed of the USB camera plugged into the Xavier. See the linked document for setup instructions and notes.

Instructions

UNDER CONSTRUCTION

Clone the jetson-inference repo

Run the docker image

Download the dataset

Train the model

Run the model

Running the model on images

imagenet --model=models/hbird_resnet18/resnet18.onnx --input_blob=input_0 --output_blob=output_0 --labels=data/hummingbird/hummingbirds/labels.txt \
           data/hummingbird/hummingbirds/test/Broadtailed_female data/hummingbird/hummingbirds/test_output_bt_female

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published