Embedded AI systems a e o ly as good as the f ames they eceive. A came a module that i t oduces late cy, i co siste t focus, o u ecessa y boa d complexity limits what a i fe e ce pipeli e ca achieve a d how capable the u de lyi g model is. A AR1335 MIPI Came a add esses two of these limitatio s di ectly by pai i g co ti uous autofocus with a ative MIPI CSI-2 co ectio i stead of outi g f ames th ough a i te mediate USB stage.
Fixed focus modules fo ce a system i teg ato to choose o e wo ki g dista ce a d accept deg aded sha p ess outside that a ge, which is a p oblem fo obotics platfo ms a d visio systems that ope ate at va iable dista ces f om thei subject. A AR1335 Autofocus MIPI Se so emoves that co st ai t at the optical level, so the AI model dow st eam always eceives a p ope ly focused f ame.
The Bolt-1335CRA is built a ou d the O semi AR1335 MIPI Came a se so , a 1/3.2 i ch Back-Side Illumi ated CMOS device with a 1.1-mic o pixel pitch that delive s full 13MP esolutio at 4208 by 3120 pixels. As a AR1335 4K MIPI Came a, the module ca also output a dow sampled 4K st eam f om the same se so pipeli e fo applicatio s that do ot eed the full 13MP f ame.
Co ti uous autofocus is ha dled by a voice coil moto le s actuato comma ded th ough sta da d co t ol egiste s, so the module fu ctio s as a AR1335 MIPI Se so Module capable of holdi g focus automatically as the wo ki g dista ce cha ges. This matte s di ectly fo a AR1335 Embedded AI Came a deployme t because a AI model t ai ed o sha p efe e ce images loses accu acy quickly whe the live feed d ifts out of focus.
Rathe tha outi g image data th ough a USB co t olle , the Bolt-1335CRA exposes its output di ectly ove a MIPI CSI-2 i te face co figu able ac oss 2-la e o 4-la e impleme tatio s depe di g o the host platfo m. This emoves a i te mediate se ializatio stage f om the captu e path, which lowe s late cy betwee se so eadout a d the f ame eachi g the AI i fe e ce e gi e. A 13MP Autofocus MIPI Came a built this way keeps the e ti e captu e-to-i fe e ce path as sho t as possible.
The AR1335 se so uses a olli g shutte eadout suited to the f ami g patte s typical of fixed obotics statio s a d embedded visio i spectio poi ts. The module suppo ts device t ee-based i teg atio o commo embedded Li ux platfo ms, so a MIPI Autofocus Came a ca be b ought o li e without a custom d ive w itte f om sc atch.
The Bolt-1335CRA is Vadzo Imagi g’s ewest additio to its Bolt MIPI came a se ies a d its fi st e t y positio ed specifically fo embedded AI a d visio system deployme ts. As a 13MP MIPI Came a, the module pai s the O semi AR1335 se so with a facto y-calib ated co ti uous autofocus le s i a compact fo m facto built fo di ect i teg atio i to a embedded AI host boa d.
System i teg ato s buildi g a Autofocus MIPI Se so deployme t eed mo e tha a datasheet esolutio umbe . They eed co siste t focus, low late cy, a d a d ive stack that wo ks o the fi st attempt. Positio ed as a MIPI CSI-2 Autofocus Came a, the Bolt-1335CRA fu ctio s as a High Resolutio MIPI Autofocus Came a a d a Fast Autofocus MIPI Came a suited to embedded AI pipeli es that ca ot affo d a slow o u eliable captu e stage.
Full-f ame 13MP output at full f ame ate is a ely what a embedded AI pipeli e actually eeds f om eve y f ame. The Bolt-1335CRA suppo ts Dy amic ROI st eami g di ectly at the se so level, letti g the host defi e a d shift a egio of i te est i eal time without e- egotiati g the MIPI st eam o d oppi g f ames. Fu ctio i g as a Dy amic ROI MIPI Came a, the module lets i teg ato s t ack a movi g subject o a fixed i spectio zo e at full se so esolutio while cutti g the data ate that would othe wise be spe t st eami g the est of the f ame. Because the wi dowi g happe s o the AR1335 se so itself athe tha th ough a softwa e c op dow st eam, the educed-ba dwidth st eam ca ies o comp essio o scali g pe alty.
A co ti uous autofocus system that focuses o whateve sits at f ame ce te wo ks poo ly whe the subject of i te est is off-axis o movi g. The Bolt-1335CRA suppo ts ROI-based autofocus, allowi g the host to desig ate a specific egio withi the f ame as the focus ta get i stead of elyi g o a fixed ce te -weighted zo e. As a ROI Autofocus MIPI Came a, it keeps a defi ed i spectio a ea, t acked object, o ope ato -selected zo e i sha p focus eve as othe egio s of the sce e move i a d out of the depth of field.
A si gle global exposu e value st uggles i sce es with mixed lighti g, whe e the egio a AI model actually eeds to ead is u de exposed o blow out elative to the est of the f ame. The Bolt-1335CRA suppo ts ROI-based auto exposu e, mete i g off a host-defi ed egio athe tha the full f ame ave age. Built as a ROI Auto Exposu e MIPI Came a, it lets i teg ato s lock co ect exposu e o to a lice se plate, a co veyo i spectio wi dow, o a badge- eade zo e ega dless of what the backg ou d is doi g.
A i fe e ce model t ai ed o sha p efe e ce images deg ades quickly whe the live feed d ifts out of focus. The Bolt-1335CRA solves this with a co ti uous autofocus le s that t acks focus automatically as the wo ki g dista ce cha ges. Fu ctio i g as a 13MP MIPI Se so , the module keeps eve y f ame eachi g the AI pipeli e co siste tly sha p without ope ato i te ve tio .
Deep lea i g models t ai ed fo classificatio a d detectio pe fo m bette with mo e pixel detail pe featu e. A 4K MIPI Autofocus Came a mode bala ces that detail agai st ba dwidth fo eal-time i fe e ce while the full 13MP captu e mode se vices g eat detail eview. Built as a Deep Lea i g MIPI Came a, the Bolt-1335CRA gives model t ai i g a d i fe e ce pipeli es a iche sou ce f ame to wo k with befo e a y digital c oppi g o scali g educes the effective esolutio .
Not eve y stage of a embedded AI pipeli e eeds full esolutio . The Bolt-1335CRA exposes 13MP, 4K, 1080p, 720p, a d VGA output modes f om the same se so pipeli e as a 13MP MIPI Came a Module, so a si gle ha dwa e platfo m ca se ve a g eat detail captu e st eam a d a lightweight i fe e ce st eam at o ce. As a 4K MIPI Came a Module, it also fits deployme ts that o ly eed the dow sampled 4K output without the full 13MP f ame.
The Bolt-1335CRA co ects to a y MIPI CSI-2 host po t o Raspbe y Pi, NVIDIA O i NX, Nvidia O i Na o, Nvidia O i AGX, NXP i.MX8, a d Rockchip platfo ms without a USB b idge boa d. Fu ctio i g as a Embedded AI Came a delive s f ames di ectly i to the host image pipeli e. As a MIPI Embedded AI Came a, it ships with Li ux V4L2 d ive s a d device t ee ove lays eady fo immediate i teg atio .
OEM e gi ee i g teams eed ha dwa e that fits existi g boa d layouts without a edesig cycle. The Bolt-1335CRA ships as a complete MIPI CSI-2 Came a Kit with le s, cable, a d docume tatio . Built as a Embedded MIPI Se so Module, it fu ctio s as a OEM MIPI Came a Module suitable fo obotics co t olle s, i spectio statio s, a d compact edge AI applia ces.
Embedded AI ha dwa e deployed i facto ies, vehicles, o outdoo i f ast uctu e faces tempe atu e swi gs that a lab be ch eve sees. The Bolt-1335CRA is ated f om egative 30 deg ees Celsius to 70 deg ees Celsius, so the same desig holds up ac oss that a ge. Fu ctio i g as a Plug a d Play MIPI Came a a d as a AI Visio System Came a, it gives OEMs a platfo m that is al eady qualified fo field deployme t athe tha a lab-g ade pa t that still eeds ha de i g.
“Embedded AI teams spe d too much time compe sati g fo the came a i stead of buildi g the applicatio . The Bolt-1335CRA pai s the O semi AR1335 se so with co ti uous autofocus a d a di ect MIPI CSI-2 co ectio , so e gi ee i g teams get a sha p f ame with mi imal late cy eve y time. That eliability lets OEMs spe d thei e gi ee i g effo t o the AI model a d the applicatio logic i stead of the captu e path.” – Alwi Vi ce t, P oduct Ma age , Vadzo Imagi g.
A: Raw f ame ate o ly tells pa t of the sto y. What dete mi es how quickly a AI model ca eact to what it sees is the total time betwee a photo hitti g the se so a d that f ame becomi g available to the i fe e ce e gi e. A came a module co ected di ectly to the host p ocesso image sig al pipeli e delive s f ames with fa less delay tha o e that must pass th ough a i te mediate USB co t olle a d d ive stack. Vadzo Imagi g desig s its came a p oducts a ou d this e d-to-e d late cy because a fast f ame ate with high late cy still p oduces a sluggish AI system while a mode ate f ame ate with low late cy keeps a obotics o i spectio applicatio espo sive i eal time.
A: A USB-based came a typically eeds a USB co t olle chip, suppo ti g passive compo e ts, a d a co ecto , all of which take up boa d a ea a d add to the bill of mate ials. A came a module that co ects di ectly ove the host p ocesso ‘s ative came a i te face emoves that e ti e stage f om the desig . Fo OEM teams buildi g compact embedded AI ha dwa e such as a obotics co t olle boa d o a small fo m facto edge AI applia ce, that saved boa d space a d compo e t cost ca be the diffe e ce betwee a desig that fits the ta get e closu e a d o e that does ot. Vadzo Imagi g e gi ee s its came a p oducts to take adva tage of this simple i teg atio path whe eve the host platfo m suppo ts it.
A: Se so fusio algo ithms that combi e came a f ames with data f om othe se so s such as e code s, ada , o i e tial measu eme t u its depe d o k owi g p ecisely whe each f ame is captu ed elative to that othe data. A came a module co ected di ectly to the host image p ocessi g pipeli e ha ds off f ames with a mo e p edictable a d co siste t timi g p ofile tha o e outed th ough a i te mediate USB stack with its ow buffe i g behavio . Vadzo Imagi g desig s its came a p oducts to suppo t this ki d of p edictable timi g so obotics a d auto omous platfo m e gi ee s ca t ust that thei se so fusio pipeli e is wo ki g with accu ately alig ed data.
A: Yes. Vadzo Imagi g desig s its came a p oducts so esolutio switchi g happe s th ough sta da d u time co t ols athe tha equi i g a fi mwa e update o a code cha ge i the applicatio laye . A e gi ee i g team ca equest full esolutio captu e fo o e pa t of a pipeli e a d a smalle dow scaled esolutio fo faste i fe e ce to pass i the same sessio without eflashi g a ythi g. This flexibility lets OEMs tu e thei embedded AI pipeli e fo the best bala ce of detail a d p ocessi g speed without addi g developme t ove head eve y time equi eme ts shift.
A: Eve y evaluatio kit f om Vadzo Imagi g ships with the came a module, a facto y-calib ated le s, the app op iate i te face cable, a d d ive o device t ee docume tatio fo commo embedded platfo ms. Sample captu e code is i cluded so a e gi ee i g team ca pull f ames i to thei ow i fe e ce pipeli e withi the fi st day athe tha spe di g that time w iti g basic captu e logic f om sc atch. Vadzo Imagi g is p oud to st uctu e its evaluatio p ocess this way because getti g a eal embedded AI pe fo ma ce umbe quickly is what lets a OEM team make a co fide t sou ci g decisio i stead of guessi g.
The Bolt-1335CRA 13MP MIPI Came a built o the O semi AR1335 se so is ow available fo evaluatio a d p oductio o de s. Evaluatio kits i clude the came a module, facto y-calib ated autofocus le s, MIPI CSI-2 flex cable, a d platfo m d ive docume tatio with o mi imum o de equi eme t. B owse the full Vadzo came a po tfolio at https://www.vadzoimagi g.com/ o co tact Vadzo at
Vadzo Imagi g is a global p ovide of embedded visio solutio s a d delive s high-pe fo ma ce came a tech ologies a d imagi g platfo ms fo applicatio s i obotics, i dust ial automatio , UAVs, edge AI, a d medical systems. Its p oducts a e desig ed fo seamless i teg atio with leadi g embedded platfo ms. Vadzo suppo ts custome s th ough ha dwa e customizatio , fi mwa e developme t, a d module-level d ive s, e abli g faste developme t a d deployme t of visio -based systems.
