Intel RealSense D435if Dimensioning Software

$250-750 USD

Posted 1 minute ago

•

Ends in 6 days

$250-750 USD

Paid on delivery

I am building a small desktop station where a fixed Intel RealSense Depth Camera D435if sits roughly 2 m above a plain, single-colour table. Each time I place an item on that table—sometimes a perfect carton, other times an irregular bag—I need the program to capture its maximum length, width, and height (in millimetres) and display the three numbers instantly on a Windows PC. Key targets • Working volume: up to 530 × 530 × 530 mm. • Accuracy: within ±5 mm, to be proven in real-world trials. • Surfaces: black, grey and multi-coloured objects must be recognised reliably. The simplest user flow is ideal: launch the app, the depth stream starts, one click (or automatic trigger) records a frame, and the bounding box results appear. A CSV or JSON export of each measurement would be a welcome extra but is not mandatory. I would like you to build it with the RealSense SDK, using Python or C++ together with OpenCV / Open3D—whichever stack you are most comfortable optimising. Please structure the code cleanly and comment it well; I intend to maintain it long term. Deliverables 1. Executable application. 2. Full, well-organised source code. 3. Step-by-step installation and calibration guide that lets me re-establish the ±5 mm accuracy after camera relocation. Once I can repeat your accuracy numbers on my side, the project will be considered successfully completed.

Project ID: 40821952

About the project

18 proposals

Open for bidding

Remote project

Active 8 secs ago

Place your bid

Bid amount Email address

Benefits of bidding on Freelancer

Set your budget and timeframe

It's free to sign up and bid on jobs

18 freelancers are bidding on average $404 USD for this job

User Avatar
Flag of PAKISTAN

Hi, I understand you need a Windows desktop app that measures an item's maximum length, width, and height from a fixed overhead depth camera, including irregular bags, and shows results instantly. I will build the measurement flow in Python or C++ with clean Software Architecture, clear calibration steps, and reliable object boundary and bounding-box handling. I will focus on the 530 mm working area, millimetre output, simple capture control, and repeatable testing toward ±5 mm. Source code, executable, setup notes, and optional CSV/JSON export will be included. Can you share the RealSense SDK version, Windows version, camera mounting details, and a few sample objects for accuracy testing? Please start the chat and let’s discuss your camera setup and test objects.

User Avatar
Flag of NIGERIA

Hi, I’m excited to propose my solution for your project. I have a decade of experience in developing similar applications, and I believe I can meet and exceed your requirements. The application will be built using the Intel RealSense SDK with Python and OpenCV for optimal performance and ease of use. The program will automatically capture the maximum length, width, and height of objects placed on the table within the specified working volume of up to 530 x 530 x 530 mm. The system will be capable of recognising black, grey, and multi-coloured objects with high accuracy, ensuring reliable measurements. To enhance user experience, the app will feature a simple user flow: launch the application, start the depth stream, take a measurement with a single click, and instantly receive the bounding box results. An optional CSV or JSON export feature will be available for each measurement. For long-term maintenance, the deliverables will include: 1. An executable application. 2. Full, well-organised source code with detailed comments. 3. A step-by-step installation and calibration guide to ensure the ±5 mm accuracy is easily re-established after camera relocation. I’ve built similar projects before, and I’m confident in meeting your accuracy standards. You can review my portfolio at https://www.freelancer.com/u/reedsystems for additional insights. Looking forward to working with you! Best regards, Reed

User Avatar
Flag of INDONESIA

Hi there, I'm excited about the opportunity to develop your dimensioning software for the Intel RealSense D435if. Your goal of instant, accurate measurement of diverse objects on a desktop station is clear, and I have hands-on experience building similar real-time vision solutions with Python, C++, OpenCV, and the RealSense SDK. I understand the importance of delivering a user-friendly Windows application that starts streaming immediately, captures one-click measurements, and delivers reliable bounding box results—while maintaining ±5 mm accuracy. Handling a variety of object shapes, sizes, and colors is a challenge I’m well-equipped for, given my background in 3D computer vision, point cloud processing, and calibration for high-accuracy measurement systems. My approach will involve robust segmentation of the object using depth data, precise extraction of the object’s maximum dimensions in all three axes, and clear on-screen display of length, width, and height. I’ll implement a streamlined interface for effortless operation, and add CSV/JSON export as an optional feature. To ensure ongoing accuracy, I’ll provide a step-by-step calibration workflow and thorough documentation so you can re-calibrate after camera relocation. The codebase will be well-organized, modular, and heavily commented for easy maintenance. Let’s work together to deliver a professional, reliable solution that meets your needs and is easy for you to use and extend in the future. Best Regards

User Avatar
Flag of VIETNAM

Hello. "INTEL REALSENSE D435IF 3D DIMENSIONING" — you need a simple Windows station that measures an object's maximum length, width, and height from a fixed overhead depth camera. I would build this around the RealSense SDK and OpenCV, using calibrated depth data, table-plane detection, object segmentation, and a 3D bounding box to calculate dimensions in millimetres. The workflow would be one-click capture with instant results, plus CSV/JSON export if useful. I would also include calibration tools and a repeatable test procedure for checking accuracy after camera relocation. The ±5 mm target needs to be validated against real objects across the full working volume, especially dark surfaces and irregular bags. I would document the test results and any measurement limitations rather than promise accuracy before testing. Can you share the table dimensions and a few representative objects, including the darkest and most irregular ones, for the initial calibration and accuracy tests? Looking forward to working with you.

User Avatar
Flag of UNITED STATES

Hi, I am Razmik from the sunny state of Florida. Your dimensioning station needs consistent 3D measurements across both regular cartons and irregular packages, with accuracy that can be verified through repeatable physical tests. I would focus on reliable depth processing, table-plane detection, and robust object segmentation rather than relying on a simple 2D bounding box. I would use the Intel RealSense SDK with Python, OpenCV, and Open3D to capture aligned depth and colour frames, calibrate the camera-to-table geometry, remove the table plane, and isolate each object. A 3D bounding-box calculation would then determine maximum length, width, and height in millimetres. The Windows interface would provide one-click capture, immediate results, and optional CSV/JSON export. I would include organized source code, a Windows executable, and a calibration guide. Accuracy would be evaluated using reference objects of different sizes, shapes, and surface colours. I would also document measurement limitations and calibration conditions rather than promise ±5 mm before real-world validation. Best regards.

User Avatar
Flag of UNITED STATES

Your station is really a repeatable measurement pipeline, not a demo. A fixed camera above a single-colour table gives a clean geometry problem, and ±5 mm is reachable if calibration is treated as a first-class feature rather than an afterthought. In Python I would pull depth and confidence frames, remove the table plane so black and grey items do not merge into the background, cluster the remaining points, then compute the minimum oriented bounding box , that is how an irregular bag returns a true maximum length, width and height instead of a loose axis-aligned guess. Distortion and scale factors live in a calibration file you can regenerate after any camera move, which is what makes your re-verification realistic. Where colour contrast is weak, an Artificial Intelligence segmentation pass tuned on your actual items recovers the outline far better than plain thresholding. The Windows front end stays thin, with plain JavaScript for the live readout, and the code splits into capture, calibration, measurement and export modules so it stays maintainable long term. What are the smallest and tallest items you will dimension, and can you confirm the camera mount is rigid and repeatable, since a ±5 mm spec at 2 m stands or falls on fixed geometry?

User Avatar
Flag of TURKEY

Hi there, I see that you require an application built with the RealSense SDK to capture the dimensions of objects accurately and display them instantly on a Windows PC. I have experience with Python and C++ and can leverage OpenCV or Open3D for optimizing the depth camera's performance in three-dimensional recognition. The focus on maintaining accuracy within ±5 mm for varied surfaces aligns with my capabilities in computer vision. I will ensure to deliver a cleanly organized source code alongside an executable application and a step-by-step guide for calibration. This will facilitate easy maintenance as you re-establish accuracy post any relocation of the camera. Looking forward to collaborating and turning your vision into reality. Regards, Ali

User Avatar
Flag of INDIA

Hello, This sounds like a precise depth-sensing application where the main challenge will be reliably detecting irregular shapes and ensuring consistent accuracy across different colors and surfaces. The RealSense SDK and OpenCV/Open3D stack should handle this well, but getting the bounding box measurements within ±5 mm for bags and cartons will require careful calibration and testing. I'd start by setting up the RealSense camera and calibrating it for the 530 × 530 × 530 mm volume, then implement a frame capture and bounding box calculation workflow. For irregular shapes, I'd use depth data to estimate dimensions, avoiding reliance on color. Once the core functionality is stable, I'd add CSV/JSON export and write the installation guide. The code will be modular and well-commented for easy maintenance. I recently worked on a similar project involving precise object measurement using depth cameras. Do you have a preference between Python or C++ for the implementation?

User Avatar
Flag of INDIA

I see you need an application to capture dimensions using the Intel RealSense D435if camera, displaying results instantly. With my experience in building computer vision applications using OpenCV and Python, I can create a user-friendly program that meets your accuracy requirements. My approach will include thorough testing to ensure that measurements fall within the ±5 mm range and that various surface colors are recognized effectively. For this project, I will start by setting up the RealSense SDK and ensuring the camera is calibrated correctly. Following that, I will implement the frame capture and measurement extraction, followed by generating a well-structured source code and user guide for long-term maintenance. You can review a similar project I completed here: Let's discuss the specifics further to ensure alignment on your requirements.

User Avatar
Flag of PAKISTAN

Hello, I can help develop a Windows desktop application that uses the Intel RealSense D435if camera to measure objects and display their maximum length, width, and height in millimetres. I would use the RealSense SDK with Python or C++, alongside OpenCV/Open3D, to capture depth frames, isolate objects from the table, process point clouds, and calculate 3D bounding dimensions. The interface would support one-click capture or automatic triggering, with optional CSV/JSON measurement export. I’ll structure the source code clearly, include installation instructions, and provide a calibration guide for camera relocation. Accuracy will be tested against physical reference measurements, with particular attention to the ±5 mm target and different object colours and shapes. Best regards, Yasir Ahmed.

内容来自 Freelancer News,版权归原作者所有。

阅读原文 ↗

把喜欢的内容,留在你的新标签页。

打开我的订阅