Best way to implement drone detection?
$begingroup$
I am new to machine learning. I am building a senior design project that utilizes three cameras and sounds an alert when a drone is seen on any of the cameras (target is a 30-50ft dome of airspace). In the end, I am going to implement a drone detection/classification model on a raspberry pi powered by an Intel Neural Compute Stick 2. I do not need to know where the drone is, I only need to know whether it is seen or not by the cameras to sound the alert.
I was unable to find any pre-trained models online that included drone detection/classification unfortunately, so unless any of you know of any I am going to create my own.
What kind of suggestions could you give for creating a model to detect if a drone is seen in the air? I started by following a binary image classification tutorial that used Keras to classify between “Santa” and “not Santa” pictures but this gave me a terribly low accuracy for the drone dataset I implemented to that tutorial (2000 “drone” and “not drone” pictures) to replace the Santa dataset (which was only around 500 pictures). It was such a bad accuracy that I put in a picture of Santa to test and it gave a 97% certainty that that picture was a “drone”.
My drone pictures were frames extracted from videos I took of 4 different kinds of drones flying at different angles and distances away and in 3 different environments. My “not drone” pictures were frames extracted from videos taken of the same environments as the “drone” pictures, but obviously without the drone flying in them.
Thank you!
machine-learning neural-network deep-learning classification
asked 2 days ago
user69202user69202
1
New contributor
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
$endgroup$
|
show 5 more comments
$begingroup$
I am new to machine learning. I am building a senior design project that utilizes three cameras and sounds an alert when a drone is seen on any of the cameras (target is a 30-50ft dome of airspace). In the end, I am going to implement a drone detection/classification model on a raspberry pi powered by an Intel Neural Compute Stick 2. I do not need to know where the drone is, I only need to know whether it is seen or not by the cameras to sound the alert.
I was unable to find any pre-trained models online that included drone detection/classification unfortunately, so unless any of you know of any I am going to create my own.
What kind of suggestions could you give for creating a model to detect if a drone is seen in the air? I started by following a binary image classification tutorial that used Keras to classify between “Santa” and “not Santa” pictures but this gave me a terribly low accuracy for the drone dataset I implemented to that tutorial (2000 “drone” and “not drone” pictures) to replace the Santa dataset (which was only around 500 pictures). It was such a bad accuracy that I put in a picture of Santa to test and it gave a 97% certainty that that picture was a “drone”.
My drone pictures were frames extracted from videos I took of 4 different kinds of drones flying at different angles and distances away and in 3 different environments. My “not drone” pictures were frames extracted from videos taken of the same environments as the “drone” pictures, but obviously without the drone flying in them.
Thank you!
machine-learning neural-network deep-learning classification
asked 2 days ago
user69202user69202
1
New contributor
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
$endgroup$
$begingroup$
Assuming you have reproduced the results for Santa data set. To check whether there is an implementation problem or not, I suggest to replace "not drone"s with "Santa"s. So 500-1000 drones, 500 Santas. If accuracy on this classification is low too, there must be an implementation problem.
$endgroup$
– Esmailian
2 days ago
$begingroup$
Do you have a class imbalance problem ? I.e.: How many frames had drones and how man did not have drones ?
$endgroup$
– Shamit Verma
2 days ago
$begingroup$
@Esmailian Okay, I will try that. Do you think I am taking the right approach for this project?
$endgroup$
– user69202
2 days ago
$begingroup$
@ShamitVerma I don’t think so. The classes are split 50-50 between the two
$endgroup$
– user69202
2 days ago
$begingroup$
@user69202 Yes your approach is correct. If the test passed, it simply means drones are harder to detect than Santas.
$endgroup$
– Esmailian
2 days ago
|
show 5 more comments
$begingroup$
I am new to machine learning. I am building a senior design project that utilizes three cameras and sounds an alert when a drone is seen on any of the cameras (target is a 30-50ft dome of airspace). In the end, I am going to implement a drone detection/classification model on a raspberry pi powered by an Intel Neural Compute Stick 2. I do not need to know where the drone is, I only need to know whether it is seen or not by the cameras to sound the alert.
I was unable to find any pre-trained models online that included drone detection/classification unfortunately, so unless any of you know of any I am going to create my own.
What kind of suggestions could you give for creating a model to detect if a drone is seen in the air? I started by following a binary image classification tutorial that used Keras to classify between “Santa” and “not Santa” pictures but this gave me a terribly low accuracy for the drone dataset I implemented to that tutorial (2000 “drone” and “not drone” pictures) to replace the Santa dataset (which was only around 500 pictures). It was such a bad accuracy that I put in a picture of Santa to test and it gave a 97% certainty that that picture was a “drone”.
My drone pictures were frames extracted from videos I took of 4 different kinds of drones flying at different angles and distances away and in 3 different environments. My “not drone” pictures were frames extracted from videos taken of the same environments as the “drone” pictures, but obviously without the drone flying in them.
Thank you!
machine-learning neural-network deep-learning classification
asked 2 days ago
user69202user69202
1
New contributor
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
$endgroup$
I am new to machine learning. I am building a senior design project that utilizes three cameras and sounds an alert when a drone is seen on any of the cameras (target is a 30-50ft dome of airspace). In the end, I am going to implement a drone detection/classification model on a raspberry pi powered by an Intel Neural Compute Stick 2. I do not need to know where the drone is, I only need to know whether it is seen or not by the cameras to sound the alert.
I was unable to find any pre-trained models online that included drone detection/classification unfortunately, so unless any of you know of any I am going to create my own.
What kind of suggestions could you give for creating a model to detect if a drone is seen in the air? I started by following a binary image classification tutorial that used Keras to classify between “Santa” and “not Santa” pictures but this gave me a terribly low accuracy for the drone dataset I implemented to that tutorial (2000 “drone” and “not drone” pictures) to replace the Santa dataset (which was only around 500 pictures). It was such a bad accuracy that I put in a picture of Santa to test and it gave a 97% certainty that that picture was a “drone”.
My drone pictures were frames extracted from videos I took of 4 different kinds of drones flying at different angles and distances away and in 3 different environments. My “not drone” pictures were frames extracted from videos taken of the same environments as the “drone” pictures, but obviously without the drone flying in them.
Thank you!
machine-learning neural-network deep-learning classification
machine-learning neural-network deep-learning classification
asked 2 days ago
user69202user69202
1
New contributor
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
asked 2 days ago
user69202user69202
1
New contributor
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
asked 2 days ago
user69202user69202
1
New contributor
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
asked 2 days ago
user69202user69202
1
asked 2 days ago
user69202user69202
1
1
New contributor
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
New contributor
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
user69202 is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
$begingroup$
Assuming you have reproduced the results for Santa data set. To check whether there is an implementation problem or not, I suggest to replace "not drone"s with "Santa"s. So 500-1000 drones, 500 Santas. If accuracy on this classification is low too, there must be an implementation problem.
$endgroup$
– Esmailian
2 days ago
$begingroup$
Do you have a class imbalance problem ? I.e.: How many frames had drones and how man did not have drones ?
$endgroup$
– Shamit Verma
2 days ago
$begingroup$
@Esmailian Okay, I will try that. Do you think I am taking the right approach for this project?
$endgroup$
– user69202
2 days ago
$begingroup$
@ShamitVerma I don’t think so. The classes are split 50-50 between the two
$endgroup$
– user69202
2 days ago
$begingroup$
@user69202 Yes your approach is correct. If the test passed, it simply means drones are harder to detect than Santas.
$endgroup$
– Esmailian
2 days ago
|
show 5 more comments
$begingroup$
Assuming you have reproduced the results for Santa data set. To check whether there is an implementation problem or not, I suggest to replace "not drone"s with "Santa"s. So 500-1000 drones, 500 Santas. If accuracy on this classification is low too, there must be an implementation problem.
$endgroup$
– Esmailian
2 days ago
$begingroup$
Do you have a class imbalance problem ? I.e.: How many frames had drones and how man did not have drones ?
$endgroup$
– Shamit Verma
2 days ago
$begingroup$
@Esmailian Okay, I will try that. Do you think I am taking the right approach for this project?
$endgroup$
– user69202
2 days ago
$begingroup$
@ShamitVerma I don’t think so. The classes are split 50-50 between the two
$endgroup$
– user69202
2 days ago
$begingroup$
@user69202 Yes your approach is correct. If the test passed, it simply means drones are harder to detect than Santas.
$endgroup$
– Esmailian
2 days ago
$begingroup$
Assuming you have reproduced the results for Santa data set. To check whether there is an implementation problem or not, I suggest to replace "not drone"s with "Santa"s. So 500-1000 drones, 500 Santas. If accuracy on this classification is low too, there must be an implementation problem.
$endgroup$
– Esmailian
2 days ago
$begingroup$
Assuming you have reproduced the results for Santa data set. To check whether there is an implementation problem or not, I suggest to replace "not drone"s with "Santa"s. So 500-1000 drones, 500 Santas. If accuracy on this classification is low too, there must be an implementation problem.
$endgroup$
– Esmailian
2 days ago
$begingroup$
Do you have a class imbalance problem ? I.e.: How many frames had drones and how man did not have drones ?
$endgroup$
– Shamit Verma
2 days ago
$begingroup$
Do you have a class imbalance problem ? I.e.: How many frames had drones and how man did not have drones ?
$endgroup$
– Shamit Verma
2 days ago
$begingroup$
@Esmailian Okay, I will try that. Do you think I am taking the right approach for this project?
$endgroup$
– user69202
2 days ago
$begingroup$
@Esmailian Okay, I will try that. Do you think I am taking the right approach for this project?
$endgroup$
– user69202
2 days ago
$begingroup$
@ShamitVerma I don’t think so. The classes are split 50-50 between the two
$endgroup$
– user69202
2 days ago
$begingroup$
@ShamitVerma I don’t think so. The classes are split 50-50 between the two
$endgroup$
– user69202
2 days ago
$begingroup$
@user69202 Yes your approach is correct. If the test passed, it simply means drones are harder to detect than Santas.
$endgroup$
– Esmailian
2 days ago
$begingroup$
@user69202 Yes your approach is correct. If the test passed, it simply means drones are harder to detect than Santas.
$endgroup$
– Esmailian
2 days ago
|
show 5 more comments
0
active
oldest
votes
Your Answer
StackExchange.ifUsing("editor", function () {
return StackExchange.using("mathjaxEditing", function () {
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
});
});
}, "mathjax-editing");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "557"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
user69202 is a new contributor. Be nice, and check out our Code of Conduct.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f46995%2fbest-way-to-implement-drone-detection%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
user69202 is a new contributor. Be nice, and check out our Code of Conduct.
user69202 is a new contributor. Be nice, and check out our Code of Conduct.
user69202 is a new contributor. Be nice, and check out our Code of Conduct.
user69202 is a new contributor. Be nice, and check out our Code of Conduct.
Thanks for contributing an answer to Data Science Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f46995%2fbest-way-to-implement-drone-detection%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
$begingroup$
Assuming you have reproduced the results for Santa data set. To check whether there is an implementation problem or not, I suggest to replace "not drone"s with "Santa"s. So 500-1000 drones, 500 Santas. If accuracy on this classification is low too, there must be an implementation problem.
$endgroup$
– Esmailian
2 days ago
$begingroup$
Do you have a class imbalance problem ? I.e.: How many frames had drones and how man did not have drones ?
$endgroup$
– Shamit Verma
2 days ago
$begingroup$
@Esmailian Okay, I will try that. Do you think I am taking the right approach for this project?
$endgroup$
– user69202
2 days ago
$begingroup$
@ShamitVerma I don’t think so. The classes are split 50-50 between the two
$endgroup$
– user69202
2 days ago
$begingroup$
@user69202 Yes your approach is correct. If the test passed, it simply means drones are harder to detect than Santas.
$endgroup$
– Esmailian
2 days ago