--- title: I Built a Stereo Vision Board for Hardware Hackers description: A hobbyist built a stereo vision board around the Rockchip RV1103, a single-core 1.1 GHz ARM Cortex-A7 SoC with 64 MB of DDR2 RAM and a MIPI CSI-2 bus in a... canonical: https://daily.dev/posts/i-built-a-stereo-vision-board-for-hardware-hackers-vih2xsfeq twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: I Built a Stereo Vision Board for Hardware Hackers | daily.dev og:description: A hobbyist built a stereo vision board around the Rockchip RV1103, a single-core 1.1 GHz ARM Cortex-A7 SoC with 64 MB of DDR2 RAM and a MIPI CSI-2 bus in a... og:url: https://daily.dev/posts/i-built-a-stereo-vision-board-for-hardware-hackers-vih2xsfeq og:image: https://api.daily.dev/og/posts/VIh2xsFEq.png og:image:alt: I Built a Stereo Vision Board for Hardware Hackers og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # I Built a Stereo Vision Board for Hardware Hackers **[Hacker News](https://daily.dev/sources/hn)** · 6 min read · 1 upvotes · 0 comments ## Summary A hobbyist built a stereo vision board around the Rockchip RV1103, a single-core 1.1 GHz ARM Cortex-A7 SoC with 64 MB of DDR2 RAM and a MIPI CSI-2 bus in a QFN-88 package. The board runs Linux built with Buildroot, using vendor BSP drivers for MIPI and hardware acceleration since not everything is upstream. Depth comes from the Rockchip Video Engine's SAD (sum of absolute differences) hardware block, matching shifted image patches from two cameras to estimate distance. The builder would like to sell a few boards if more than 20 people sign up for email notifications, and hopes to later add SLAM and YOLO support if there is enough interest. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is the Rockchip RV1103 SoC and what are its specs? The Rockchip RV1103 is a single-core 1.1 GHz ARM Cortex-A7 system on chip, packaged in a QFN-88 footprint. It includes 64 MB of DDR2 RAM and a MIPI CSI-2 camera bus, making it suitable for small embedded vision boards that need built-in compute and Linux support. _Hobbyists prototyping SoC-based vision boards can track RV1103 projects and builds on daily.dev._ ### How does SAD-based stereo block matching estimate depth from two cameras? It compares vertical columns of image patches from two cameras separated by a horizontal baseline distance, shifting one image relative to the other using sum of absolute differences (SAD). When patches match at an early shift the object is far away; a later shift (more shifting needed) indicates the object is closer, so the shift amount at the match becomes the depth value. _Developers implementing stereo depth algorithms can follow practical SAD matching write-ups on daily.dev._ ### What is the boot sequence for Rockchip ARM SoCs running Linux with U-Boot? It goes BootROM (loads TPL from ROM using SRAM), then TPL (Tertiary Program Loader, a Rockchip binary that sets up the DRAM controller), then SPL (Secondary Program Loader, bundled with U-Boot) which loads U-Boot, which then loads the Linux kernel, followed by an init system starting services. _Engineers bringing up custom ARM boards can compare bootloader stage breakdowns like this on daily.dev._ ## Community take How the wider developer community reacted, aggregated from 1 discussion and 15 comments across hackernews (as of 2026-10-25). **TL;DR:** Interest centers on the technical details of the SAD stereo-matching approach, with experienced commenters pointing out that calibration is the hard part the project seems to gloss over, while others are simply impressed and curious about cost. **Sentiment:** 25% positive · 55% mixed · 20% skeptical **The case for** - Some find the project neat and would want one if they had a use case. - SAD matching is noted to run efficiently, even on modest hardware like a Pi. - Stereo disparity math conveniently limits the search range needed, making the approach efficient. **The pushback** - Several suspect the demo wasn't properly calibrated and may be running on raw, unrectified images. - Calibration (intrinsics/extrinsics) is seen as the hard, glossed-over part of stereo vision that the post doesn't address. - Stereo matching is noted to be sensitive to repeating textures, which could cause misjudged distances. **By community** - hackernews (mixed): Technical commenters dig into calibration, algorithm tradeoffs, and alternatives like structure-from-motion, with a mix of praise and skepticism about demo quality. **Hottest debate:** Whether the demo actually involved proper stereo calibration or just raw unrectified images. **Open questions** - What is the bill-of-materials cost for the board? - How does the system perform when tested on a second hardware platform? **Highlights** > I'm not convinced he's done the calibration, at least in his demo video. It's pretty bad, imo. I think he's just running it on the raw unrectified images. For anyone wanting to play with this, OpenCV has a number of tutorials on this using stuff like stereoBM (which also uses SAD if I'm remembering my algos). The first step is calibration -- intrinsics (basically figuring out the equations that can cancel out the lens distortion) and extrinsics (how far apart are the two cameras in XYZ, tilt, etc). > — [hex4def6 on hackernews · 1 comments](https://news.ycombinator.com/item?id=50228606) > SAD runs in better than real time on a typical desktop CPU, even on a Pi. The author glossed over basically all of the calibration involved with stereo. It’s not quite as simple as taking two cameras and running a matcher. Because stereo disparity (parallax) scales inversely with distance, conveniently you don’t need to search huge ranges. A point at infinity should have zero disparity in a calibrated stereo system. So normally you pick a depth range and search the corresponding pixels. Also if the cameras are calibrated, you only need to search horizontally so it works out to say 64 operations per pixel, which can be vectorized if the CPU supports it. Other approaches use more stable comparisons like census (is each pixel in the match window brighter than the central one) and hamming distance which is also very efficient. And yes, stereo is sensitive to repeating texture but absence of texture like a blank wall is usually worse. > — [throwaway219450 on hackernews · 2 comments](https://news.ycombinator.com/item?id=50228336) > The SAD algorithm looks neat! How long does it take? Wouldn’t there be a large range of offsets? Or does limiting the minimum distance help mitigate such? I suppose there could be corner cases in which an object with regular pattern could be misinterpreted as being far away ? > — [BobbyTables2 on hackernews · 1 comments](https://news.ycombinator.com/item?id=50228143) > You can use structure from motion for this. It’s the same way that drone photogrammetry works, ie bundle adjustment. It’ll give you the intrinsics at least. > — [throwaway219450 on hackernews](https://news.ycombinator.com/item?id=50228877) **Source threads** - [hackernews](https://news.ycombinator.com/item?id=50172432) · 79 points · 15 comments ## Similar posts on daily.dev - [A Stereo Vision Board For Hardware Hackers](https://daily.dev/posts/a-stereo-vision-board-for-hardware-hackers-mxxisfjo5) · Hackaday · 0 upvotes · 0 comments - [A modular USB, HDMI, and Ethernet camera system \(Crowdfunding\)](https://daily.dev/posts/a-modular-usb-hdmi-and-ethernet-camera-system-crowdfunding--wysynkmun) · CNX Software · 0 upvotes · 0 comments - [This Tiny SBC Lets You Swap Interface Modules for Voice, Vision or Sensing](https://daily.dev/posts/this-tiny-sbc-lets-you-swap-interface-modules-for-voice-vision-or-sensing-scmsawzzq) · It's Foss · 0 upvotes · 0 comments --- Tags: [#hardware](https://daily.dev/tags/hardware), [#computer-vision](https://daily.dev/tags/computer-vision), [#embedded](https://daily.dev/tags/embedded) [View this post on daily.dev](https://daily.dev/posts/i-built-a-stereo-vision-board-for-hardware-hackers-vih2xsfeq) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"I Built a Stereo Vision Board for Hardware Hackers","url":"https://daily.dev/posts/i-built-a-stereo-vision-board-for-hardware-hackers-vih2xsfeq","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/i-built-a-stereo-vision-board-for-hardware-hackers-vih2xsfeq"},"datePublished":"2026-10-25T03:16:01.467Z","dateModified":"2026-10-25T10:01:47.603Z","description":"A hobbyist built a stereo vision board around the Rockchip RV1103, a single-core 1.1 GHz ARM Cortex-A7 SoC with 64 MB of DDR2 RAM and a MIPI CSI-2 bus in a...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/aad1519cda6c8104c5a16b22e836e563?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/aad1519cda6c8104c5a16b22e836e563?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Hacker News","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Hacker News","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/hn","url":"https://daily.dev/sources/hn"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/i-built-a-stereo-vision-board-for-hardware-hackers-vih2xsfeq","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"hardware,computer-vision,embedded","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Hacker News","item":"https://daily.dev/sources/hn"},{"@type":"ListItem","position":3,"name":"I Built a Stereo Vision Board for Hardware Hackers"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/i-built-a-stereo-vision-board-for-hardware-hackers-vih2xsfeq#faq","mainEntity":[{"@type":"Question","name":"What is the Rockchip RV1103 SoC and what are its specs?","acceptedAnswer":{"@type":"Answer","text":"The Rockchip RV1103 is a single-core 1.1 GHz ARM Cortex-A7 system on chip, packaged in a QFN-88 footprint. It includes 64 MB of DDR2 RAM and a MIPI CSI-2 camera bus, making it suitable for small embedded vision boards that need built-in compute and Linux support. Hobbyists prototyping SoC-based vision boards can track RV1103 projects and builds on daily.dev."}},{"@type":"Question","name":"How does SAD-based stereo block matching estimate depth from two cameras?","acceptedAnswer":{"@type":"Answer","text":"It compares vertical columns of image patches from two cameras separated by a horizontal baseline distance, shifting one image relative to the other using sum of absolute differences (SAD). When patches match at an early shift the object is far away; a later shift (more shifting needed) indicates the object is closer, so the shift amount at the match becomes the depth value. Developers implementing stereo depth algorithms can follow practical SAD matching write-ups on daily.dev."}},{"@type":"Question","name":"What is the boot sequence for Rockchip ARM SoCs running Linux with U-Boot?","acceptedAnswer":{"@type":"Answer","text":"It goes BootROM (loads TPL from ROM using SRAM), then TPL (Tertiary Program Loader, a Rockchip binary that sets up the DRAM controller), then SPL (Secondary Program Loader, bundled with U-Boot) which loads U-Boot, which then loads the Linux kernel, followed by an init system starting services. Engineers bringing up custom ARM boards can compare bootloader stage breakdowns like this on daily.dev."}}]} ```