
Closed
Posted
Paid on delivery
I need a Windows desktop application built with Python 3.12 that connects to a Hikrobot industrial camera through the MVS SDK, runs a custom YOLO (Ultralytics) model via OpenCV/NumPy, and reliably detects individual metal bars scattered on a work surface. Any orientation—horizontal, vertical, diagonal, or arbitrarily rotated—must be handled, and when bars overlap the software should still outline each one separately and decide which bar sits physically on top by analysing the single 2-D image alone (no depth sensors or special labels). While the camera is streaming the GUI should: • show the live image with coloured outlines around every detected bar, each overlay carrying a unique ID number; • display, in a side panel or table, the ID, 2-D position, rotation angle, and a “Top / Below” indicator that states whether that bar is above or underneath another. Deliverables 1. Executable Windows desktop application and all source code. 2. Trained YOLO weights (with brief notes on the training dataset so I can retrain later). 3. Setup and run instructions that cover Python 3.12 environment creation, required pip packages, and Hikrobot MVS SDK integration. 4. A short test video or image set demonstrating correct detection, orientation output, and top-bar identification. Acceptance criteria • Detection accuracy remains stable when bars are rotated to any angle or partially occlude one another. • GUI refreshes in near real-time on a typical industrial PC. • Top/Below flag matches visual inspection on at least 95 % of overlapping cases in the provided test set.
Project ID: 40787674
51 proposals
Remote project
Active 2 days ago
Set your budget and timeframe
Get paid for your work
Outline your proposal
It's free to sign up and bid on jobs
51 freelancers are bidding on average ₹7,270 INR for this job

Hello, I trust you're doing well. I am well experienced in machine learning algorithms, with nearly a decade of hands-on practice. My expertise lies in developing various artificial intelligence algorithms, including the one you require, using Python, and similar tools. I have worked with pytorch, and tensorflow to develop DL models, .I hold a doctorate from Tohoku University and have a number of publications in the same subject. My portfolio, which showcases my past work, is available for your review. Your project piqued my interest, and I would be delighted to be part of it. Let's connect to discuss in detail. Warm regards.
₹7,000 INR in 7 days
7.3
7.3

Overlapping bars are the hard part here. A plain YOLO box can't tell you which bar is on top. But a segmentation mask can: where two bars cross, the top bar's outline runs unbroken and the lower one gets cut. That's the rule I'd train and test for. I built a Brinell hardness analyzer for a materials testing lab. It grabs frames from an industrial camera and measures the indent with OpenCV from a single 2D image. So camera capture plus precise measuring is work I've done before. For your app I'll use Python 3.12 with the Hikrobot MVS SDK to grab frames. An Ultralytics YOLO segmentation model finds each bar, and OpenCV gets its centre and angle from the mask. The GUI shows the live feed with coloured outlines and IDs, plus a table with ID, position, angle and Top / Below. Detection runs in its own thread, so the display stays smooth. You'll get the exe, full source, the weights, dataset notes for retraining, setup steps and a test video. About one week once I have sample images. How many real images of the bars can you capture from your actual camera setup? Labelled overlap shots under your own lighting will decide that 95% far more than the model choice. Haleem
₹7,000 INR in 7 days
5.9
5.9

Dear Hiring Manager, Thanks for the job posting and hope you are doing great! I am a Computers grad with 21 years experience in this industry in companies like Microsoft. I have done many such projects in the profile here and am rated all 5-stars :) If you want, I can do it for you. More once you revert. Regards, Anurag.
₹8,000 INR in 7 days
6.1
6.1

The difficult part isn’t detecting rotated bars, it’s separating overlapping instances and inferring which one is on top from 2-D appearance alone. I’d use YOLO segmentation or OBB rather than plain boxes, derive each bar’s centroid/angle from its mask, and keep tracking IDs stable across frames. For overlaps, I’d classify depth order from occlusion boundaries and visible-edge continuity, but the 95% target must be validated against your actual surface, lighting and bar geometry. MVS acquisition and inference would run off the GUI thread to keep live refresh responsive.
₹1,500 INR in 7 days
5.7
5.7

Standard YOLO bounding boxes will fail here. Diagonal bars produce boxes that overlap heavily, so NMS merges them and your Top/Below logic has nothing clean to work from. Ultralytics YOLO segmentation or OBB is the fix. Quick questions - what Hikrobot model, resolution, and max bar count are we dealing with? And does the industrial PC have an NVIDIA GPU, or must inference run CPU-only via ONNX? Here is the architectural approach: - COMPUTER VISION: Train Ultralytics YOLO-seg on rotated, overlapping bars so each one gets its own mask, ID, and OpenCV minAreaRect angle. - IMAGE PROCESSING: Decide Top/Below by checking which mask keeps unbroken edges at each intersection, validated against your 95% target. - SOFTWARE ARCHITECTURE: Run MVS SDK capture and inference on worker threads so the Python
₹6,300 INR in 3 days
5.8
5.8

Hello, Problem: Manual object detection in images or video frames is time-consuming, and users need a simple interface to analyze visual data efficiently. Solution: I will develop a user-friendly Python GUI using Tkinter or PyQt with OpenCV and a suitable object-detection model. The application will support image or video input, display detected objects with bounding boxes and confidence scores, and provide configurable detection settings and exportable results. I will also include error handling and clear usage instructions. Result: You will receive a practical object-detection tool with an intuitive interface, reliable processing, and maintainable source code. With 20 years of software development experience, I focus on robust automation tools and maintainable applications. 1. Should the GUI process images, videos, or both? 2. Do you require a specific object-detection model? Thanks
₹12,000 INR in 7 days
5.7
5.7

Python, machine learning, and software architecture are the strongest match for building your Windows object-detection GUI. I can develop the application around your custom YOLO model, OpenCV/NumPy processing, and Hikrobot camera integration through the MVS SDK, with a clear interface for live imagery, detection IDs, positions, rotation angles, and Top/Below status. My C++ programming background is also relevant to industrial-camera environments and SDK-based integration, while my software architecture skills support a maintainable solution that includes the executable, source code, model weights, dataset notes, setup instructions, and demonstration test media. I understand that arbitrary bar rotation and overlap require more than basic bounding-box detection, so the application must be structured around object separation and image-based occlusion analysis rather than assuming a fixed orientation. I would be glad to apply my Python and ML experience to deliver this desktop vision system and discuss the project further.
₹5,000 INR in 5 days
5.2
5.2

Separating overlapping metal bars and deciding which one sits on top from a single 2-D frame is the interesting part here, and an oriented detection model makes it much more reliable. My approach is a YOLO OBB (oriented bounding box) or segmentation model in Ultralytics, trained on images captured from your Hikrobot camera, so each bar gets its own rotated outline and angle. For the Top/Below flag I would analyse the overlap regions using mask occlusion and edge continuity, since the bar on top keeps an unbroken contour. The GUI would be PySide6 with the MVS SDK feed running in a worker thread, live overlays with IDs, and a side table for position, angle and status, packaged as a Windows executable. I have 9 years of full-stack and Python experience, including AI/ML work, and I deliver clean source code with setup notes so you can retrain later. Do you already have labelled images of the bars, or should dataset capture and labelling be part of this? Which camera model and resolution are you using? This amount covers a first working phase, and I am happy to confirm the full scope in a chat here on Freelancer.
₹12,000 INR in 14 days
4.4
4.4

Hi, there I successfully developed a custom YOLO model-based object detection system for a manufacturing client using OpenCV/NumPy with Ultralytics YOLOv7 model, achieving 99% detection accuracy on variously oriented metal parts in 2D images. The system handled arbitrarily rotated objects and partially occluded cases, enabling accurate top/below flag assignment. I implemented features like real-time GUI updates, live image streaming, and coloured outlines around detected objects, each with a unique ID. I also developed a setup script to integrate the Hikrobot MVS SDK with Python 3.12 environment. For this project, I plan to leverage my experience with custom YOLO model training, OpenCV image processing, and MVS SDK integration to deliver a robust and accurate object detection system. I will use Ultralytics YOLOv7 model, fine-tuned for metal bar detection, and implement a real-time GUI with live image streaming, coloured outlines, and accurate top/below flag assignment. I will also provide detailed setup instructions and a test video demonstrating the system's capabilities. If I use my previous experience, your project will likely be completed successfully. Hope to discuss this in detail. Through detailed discussion, I think I can find the better solution to finish your project successfully. Thank you.
₹7,000 INR in 7 days
2.7
2.7

Hi, I recently worked on a highly advanced computer vision project involving multi-drone detection and tracking using YOLO, OpenCV, and object-tracking pipelines. This experience gave me hands-on exposure to real-time object detection, multi-object tracking, and challenging visual scenarios, making this project a great fit for my expertise. I can develop your Windows desktop application using Python 3.12, Hikrobot MVS SDK, Ultralytics YOLO, OpenCV, and NumPy. My approach includes: • Integrating the Hikrobot camera for reliable live streaming. • Detecting metal bars at arbitrary orientations using oriented bounding boxes and instance segmentation. • Separately identifying partially overlapping bars wherever sufficient visual information is available. • Displaying colour-coded outlines, unique IDs, 2D coordinates, rotation angles, and Top/Below status in a real-time GUI. • Optimising inference and GUI performance for near-real-time operation. • Delivering the executable, complete source code, trained weights, dataset documentation, setup instructions, and test demonstrations. For determining which bar is physically on top, I would analyse occlusion boundaries and visible contours using 2D image processing. Let’s discuss your camera model and available dataset to determine the best implementation strategy. Best regards
₹7,500 INR in 7 days
2.8
2.8

Hello, I understand you need a Python 3.12 Windows application integrating a Hikrobot camera, YOLO detection, and a real-time GUI to identify and track metal bars at different orientations. Here’s what I can provide: 1. Integrate the Hikrobot MVS SDK with OpenCV, NumPy, and Ultralytics YOLO for live camera capture and rotated-bar detection. 2. Build a desktop GUI displaying colour-coded outlines, unique IDs, 2D positions, rotation angles, and a Top/Below status table. 3. Package the source code, executable, model weights, setup instructions, and test examples, with performance and accuracy validation on your dataset. I bring over 4+ years of experience in Python, computer vision, API integration, and software development, focusing on reliable, maintainable applications. Just to clarify a few things: 1. Do you have labelled images and trained YOLO weights, or should model training and annotation be included? 2. Can you provide overlapping-bar test images and confirm the Hikrobot camera model and industrial PC specifications? Please come to the chat box to discuss more about your project. Best regards Indresh Kushwaha
₹7,000 INR in 7 days
1.7
1.7

Hi, I'll approach your metal-bar detection system by combining YOLO-based instance segmentation, OpenCV geometry analysis, and a responsive Python desktop interface. I'll begin by reviewing sample images and the Hikrobot MVS camera setup to identify the best approach for detecting rotated and overlapping bars. Next, I'll train or fine-tune a YOLO segmentation model, using each bar's mask to calculate its position, orientation angle, and unique detection ID. For overlapping bars, I'll develop an image-based relationship classifier using visible contours, occlusion boundaries, and labeled training examples to estimate which bar is above or below. I'll integrate the detection pipeline with the Hikrobot MVS SDK and build a Windows GUI displaying live video overlays and a table of measurements. My experience includes Python, OpenCV, machine learning integration, application development, and software debugging. I'll validate detection speed and accuracy against your test images before packaging the Windows application, source code, and documentation. Could you provide sample overlapping-bar images and confirm the Hikrobot camera model? Best regards, Braian Alves
₹5,000 INR in 15 days
1.2
1.2

Two bars crossing in an X can share almost the same upright box, and a standard YOLO detector's overlap filter then keeps one and drops the other, so the hidden bar never gets an ID. Your any-angle rule needs rotated boxes. Ultralytics' OBB models return each bar's centre, size and rotation directly, which fills your position and angle columns, and overlap is judged on the rotated boxes, so both crossing bars survive. Labels matter too: a bar half under another gets one full-length box, or the model learns it as two short bars with two IDs. Top or below: where two rotated boxes overlap, the top bar's edges run unbroken across the crossing, while the lower bar's edges stop at the top bar's sides. A small second model reads each crossing patch, turned so one bar lies flat, and learns that cue from your labelled piles. It's checked against your 95% rule on the test set before handover. The camera runs through Hikrobot's MVS Python wrapper on its own capture thread, so a slow frame never freezes the GUI. PySide6 shows the live overlay with IDs and the table beside it, packaged as a Windows exe with weights, dataset notes and Python 3.12 setup steps. I've built real-time object detection and tracking products and edge AI on Jetson, in Python. Hikrobot's SDK is newer to me. My rating here is 5 out of 5. ₹4,200, 5 days. A bar in a pile of three can lie on one bar and under another. Should the table flag it Below, or show its flag against each bar it crosses? Ronak
₹4,200 INR in 5 days
1.1
1.1

Hello Mate!Greetings , Good evening! I am skilled mobile coder with skills including C++ Programming, YOLO, Windows Desktop, OpenCV, Python, Computer Vision, Software Architecture, Machine Learning (ML), NumPy and Image Processing. Please send a message to discuss more regarding this project. Thanks
₹7,770 INR in 6 days
0.0
0.0

The object detection GUI starts with a Python 3.12 prototype using recorded camera images to validate rotated outlines and overlap reasoning. This milestone buys image loading, bar IDs, positions, angles and a preliminary Top/Below table, with a reviewable checkpoint within three working days after funding. I’ll use instance segmentation for separate outlines and mark ambiguous overlaps for review rather than assign unsupported depth. The checkpoint will measure detection and overlap classification against your labelled examples, assessing feasibility of the 95% target. A next milestone covers MVS streaming, model training, performance tuning, Windows packaging, source code, setup instructions and demonstration outputs. Please send representative camera images showing overlapping bars, their expected top/below ordering, and any existing YOLO weights. Produced using AI tools and checked before delivery. Price: 11825 INR fixed · Timeline: 3 working days from access/brief Konrad
₹11,825 INR in 3 days
0.0
0.0

I see that you're looking for a Windows desktop application to detect metal bars using a Hikrobot industrial camera and a custom YOLO model. My strong background in Python, OpenCV, and machine learning positions me well to deliver this application robustly. I will first analyze the requirements thoroughly to ensure accurate detection of bars in all orientations while managing overlaps effectively. This will involve creating the GUI to show live images with unique ID outlines and a detailed side panel. I have worked on AI-powered applications, as seen in my project on building an AI music quiz platform. You can check it out here: My approach will ensure high accuracy and a responsive GUI on typical industrial PCs. I look forward to discussing how I can bring this project to life.
₹1,500 INR in 3 days
0.0
0.0

Hi, I can develop this as a Python 3.12 Windows application integrating the Hikrobot MVS SDK, Ultralytics YOLO, OpenCV and NumPy. The application can include: • Live Hikrobot camera acquisition • Custom YOLO model for individual metal-bar detection • Detection across arbitrary rotations • Separate IDs and contours for detected bars • 2D position and rotation-angle calculation • Occlusion/overlap analysis for Top/Below classification • Near-real-time annotated camera preview • Side-panel table with ID, coordinates, angle and stacking status • Clean Windows GUI and packaged executable • Full source, trained weights and retraining notes • Environment/MVS SDK setup documentation • Test images/video demonstrating results The Top/Below requirement is the key technical challenge because it must be inferred from a single 2D image. I would validate that logic against your actual bar appearance and dataset rather than claim 95% accuracy before testing. Please share sample camera images/video, bar types, expected working distance, image resolution, and any existing labelled dataset. Looking forward to hearing from you.
₹5,000 INR in 7 days
0.0
0.0

I can deliver 'Object Detection GUI' end to end — clean architecture, tests, and deployment. I'll start with a short plan, share progress daily, and hand over production-ready code. Could you confirm your timeline and must-have features so I can lock the scope?
₹6,661.28 INR in 7 days
0.0
0.0

Efficient detection of individual metal bars in arbitrary orientations requires robust modeling and effective real-time processing. I'll build a Windows desktop application in Python 3.12 that uses the Hikrobot MVS SDK for camera integration and employs an optimized YOLO model via OpenCV/NumPy for precise detection. I'll implement image processing techniques to handle overlapping bars, ensuring accurate ID assignment and a reliable "Top / Below" indicator through contour analysis. The GUI will display live image feeds with clear overlays of detected bars, dynamically updating side panels with ID, position, rotation, and hierarchy. I'll provide a comprehensive package of deliverables, including an executable application, trained YOLO weights with retraining notes, setup instructions for the Python environment and SDK integration, and a short video demonstrating the system’s capabilities. What specific performance benchmarks or additional features do you envision for the application? Liam
₹5,000 INR in 7 days
0.0
0.0

Building a reliable Windows desktop vision application is well aligned with my 6+ years of software engineering experience and core strengths in Python, NumPy, software architecture, and Windows Desktop development. I can create a structured application that connects the Hikrobot camera workflow with the custom YOLO/Ultralytics model and presents detections clearly in a live GUI. My Python and NumPy experience supports the image-processing and coordinate calculations needed for rotated bars, while my software architecture background helps keep camera streaming, inference, tracking IDs, overlays, and the side-panel data organised and maintainable. I can also deliver the source code, executable application, trained weights with dataset notes, environment and MVS SDK setup instructions, and a test video or image set covering orientation and overlap cases. I would be glad to develop this object-detection GUI as a robust Windows application and provide a complete, usable deliverable.
₹10,000 INR in 7 days
0.0
0.0

New Delhi, India
Member since Oct 26, 2026
₹12500-37500 INR
£20-250 GBP
₹12500-37500 INR
₹500000-1000000 INR
₹12500-37500 INR
$30-250 USD
$30-250 USD
₹7000-10000 INR / hour
₹12500-37500 INR
₹12500-37500 INR
$250-750 USD
$30-250 CAD
₹750-1250 INR / hour
$15-25 USD / hour
₹1500-12500 INR
₹1500-12500 INR
$250-750 USD
$15-25 USD / hour
₹1500-12500 INR
€30-250 EUR