{"id":7717,"date":"2026-10-23T14:54:07","date_gmt":"2026-10-23T06:54:07","guid":{"rendered":"https:\/\/www.unitecho.net\/?p=7717"},"modified":"2026-11-06T10:15:58","modified_gmt":"2026-11-06T02:15:58","slug":"how-does-an-optical-sorting-machine-work-camera-lighting-and-ejection-explained","status":"publish","type":"post","link":"https:\/\/www.unitecho.net\/fr\/knowledge-news\/how-does-an-optical-sorting-machine-work-camera-lighting-and-ejection-explained\/","title":{"rendered":"Comment fonctionne une trieuse optique ? Explications sur la cam\u00e9ra, l'\u00e9clairage et l'\u00e9jection"},"content":{"rendered":"

Le contr\u00f4le qualit\u00e9 constitue d\u00e9sormais le principal goulot d\u2019\u00e9tranglement sur de nombreuses lignes de production \u00e0 haut d\u00e9bit. Une usine traitant plusieurs tonnes par heure ne peut pas affecter suffisamment de personnel pour examiner chaque pi\u00e8ce, et le jugement humain est sujet \u00e0 des variations au cours d\u2019un poste de travail. Les acheteurs ont renforc\u00e9 leurs sp\u00e9cifications en mati\u00e8re de corps \u00e9trangers, d\u2019uniformit\u00e9 des couleurs et de pr\u00e9cision dimensionnelle, tandis que les march\u00e9s des mati\u00e8res secondaires accordent une prime \u00e0 la puret\u00e9. Le tri optique est pass\u00e9 du statut d\u2019am\u00e9lioration de la qualit\u00e9 \u00e0 celui d\u2019op\u00e9ration standard.<\/p>\n\n\n\n

Un machine de tri optique<\/a> Le syst\u00e8me fonctionne en r\u00e9partissant les produits en une seule couche, en les acheminant \u00e0 travers une zone d'inspection o\u00f9 un \u00e9clairage contr\u00f4l\u00e9 et des cam\u00e9ras ou des capteurs spectraux capturent une image de chaque article, en comparant chaque image en temps r\u00e9el \u00e0 des crit\u00e8res d'acceptation et de rejet, et en d\u00e9clenchant un \u00e9jecteur pneumatique ou m\u00e9canique d\u00e8s qu'un article rejet\u00e9 atteint le point de s\u00e9paration. Le cycle complet, de la lumi\u00e8re r\u00e9fl\u00e9chie \u00e0 l'impulsion d'air, s'effectue g\u00e9n\u00e9ralement en quelques dizaines de millisecondes.<\/strong><\/p>\n\n\n\n

Ce qui distingue un trieur performant d'un trieur qui se contente de fonctionner, c'est l'\u00e9quilibre entre trois sous-syst\u00e8mes.<\/p>\n\n\n\n

\"machine<\/figure>\n\n\n\n

Comment fonctionne r\u00e9ellement une machine de tri optique ?<\/h2>\n\n\n\n

Une trieuse optique est un syst\u00e8me automatis\u00e9 d'inspection et de tri qui \u00e9value chaque article d'un flux en mouvement selon des crit\u00e8res programm\u00e9s et \u00e9limine ceux qui ne r\u00e9pondent pas aux crit\u00e8res, sans toucher \u00e0 ceux qui les satisfont.<\/strong><\/p>\n\n\n\n

Les tamis, les aimants et les tables de densit\u00e9 effectuent un tri en fonction d\u2019une propri\u00e9t\u00e9 globale et ne permettent de diviser un flux qu\u2019en fractions grossi\u00e8res. Une machine de tri optique, en revanche, \u00e9value chaque objet individuellement : teinte incorrecte, forme incorrecte, contamination, \u00e9br\u00e9chure ou polym\u00e8re inadapt\u00e9. L\u2019inspection \u00e9tant non destructive, elle fonctionne \u00e0 plein d\u00e9bit tout en examinant l\u2019ensemble de la production plut\u00f4t qu\u2019un simple \u00e9chantillon.<\/p>\n\n\n\n

Les quatre sous-syst\u00e8mes et les raisons pour lesquelles ils ne peuvent pas \u00eatre dissoci\u00e9s<\/h3>\n\n\n\n

Chaque plateforme, qu'il s'agisse d'un trieur \u00e0 goulottes compact ou d'une machine \u00e0 large bande, comprend les quatre m\u00eames modules.<\/p>\n\n\n\n

Sous-syst\u00e8me<\/td>Fonction<\/td>Sympt\u00f4me typique de d\u00e9faillance<\/td><\/tr>
Flux<\/td>\u00c9tale le mat\u00e9riau en une monocouche \u00e0 vitesse constante<\/td>Le chevauchement masque les d\u00e9fauts ; un espacement irr\u00e9gulier perturbe la synchronisation<\/td><\/tr>
Optique<\/td>Met en \u00e9vidence le flux et affiche une image pour chaque \u00e9l\u00e9ment<\/td>Faible contraste, flou de mouvement, vitres sales<\/td><\/tr>
Traitement<\/td>Effectue une classification par rapport \u00e0 des seuils ou \u00e0 un mod\u00e8le entra\u00een\u00e9<\/td>Variations des seuils d'instabilit\u00e9 d'un lot \u00e0 l'autre<\/td><\/tr>
S\u00e9paration<\/td>D\u00e9tourne les produits rejet\u00e9s hors du flux principal<\/td>Pouls tardif ou faible, r\u00e9manence, omissions<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Les modules sont interd\u00e9pendants, ce qui p\u00e9nalise les mises \u00e0 niveau partielles. Une cam\u00e9ra haute r\u00e9solution plac\u00e9e derri\u00e8re une source lumineuse mal choisie produit une image nette d'un d\u00e9faut qu'elle ne peut distinguer de l'arri\u00e8re-plan, et une classification correcte transmise avec trois millisecondes de retard parvient \u00e0 l'\u00e9jecteur une fois que la cible est d\u00e9j\u00e0 pass\u00e9e.<\/p>\n\n\n\n

Comment fonctionne le tri optique, \u00e9tape par \u00e9tape<\/h2>\n\n\n\n

Le tri optique s'effectue selon un cycle continu en quatre \u00e9tapes : s\u00e9paration des produits, \u00e9clairage et acquisition d'images, classification, puis d\u00e9viation des produits rejet\u00e9s. Chaque \u00e9tape est synchronis\u00e9e avec la vitesse de transport afin que la d\u00e9cision et l'action physique s'appliquent au m\u00eame objet.<\/strong><\/p>\n\n\n\n

Cette architecture est d\u00e9crite dans la documentation de r\u00e9f\u00e9rence disponible sur tri optique<\/a>, et la s\u00e9quence ci-dessous la reprend en mettant l'accent sur les points o\u00f9 la pr\u00e9cision est effectivement perdue.<\/p>\n\n\n\n

\u00c9tape 1 : Alimentation et s\u00e9paration des pi\u00e8ces<\/h3>\n\n\n\n

Le produit passe par une tr\u00e9mie et est r\u00e9parti par un alimentateur vibrant, un tapis de nivellement ou un convoyeur d'acc\u00e9l\u00e9ration jusqu'\u00e0 ce que les pi\u00e8ces circulent en une seule couche, \u00e0 un espacement r\u00e9gulier. Deux pi\u00e8ces qui se chevauchent forment une seule silhouette ; ainsi, un d\u00e9faut sur la face cach\u00e9e est invisible. Un espacement irr\u00e9gulier est tout aussi pr\u00e9judiciable, car l\u2019\u00e9jecteur est constitu\u00e9 d\u2019un ensemble de buses fixes et un article arrivant en retard re\u00e7oit une impulsion dirig\u00e9e vers l\u2019endroit o\u00f9 il aurait d\u00fb se trouver.<\/p>\n\n\n\n

\u00c9tape 2 : D\u00e9tection dans la zone d'inspection<\/h3>\n\n\n\n

Des sources lumineuses \u00e9clairent chaque \u00e9chantillon et des cam\u00e9ras captent la lumi\u00e8re r\u00e9fl\u00e9chie, transmise ou \u00e9mise. La r\u00e9flectance permet de d\u00e9terminer la couleur, la texture et la forme. La transmission, g\u00e9n\u00e9r\u00e9e par un r\u00e9tro\u00e9clairage, permet de distinguer la silhouette et la structure interne. L'\u00e9mission, g\u00e9n\u00e9ralement sous forme de fluorescence provoqu\u00e9e par une excitation aux ultraviolets, permet d'identifier la nature du mat\u00e9riau pour certains contaminants sp\u00e9cifiques.<\/p>\n\n\n\n

\u00c9tape 3 : Classification et fen\u00eatre de d\u00e9cision<\/h3>\n\n\n\n

Le processeur extrait de chaque objet la teinte, la surface, le contour, le rapport d'aspect, la texture et la r\u00e9ponse spectrale, puis compare ce vecteur \u00e0 des seuils d\u00e9finis par l'op\u00e9rateur ou \u00e0 un mod\u00e8le entra\u00een\u00e9. L'\u00e9ventail des caract\u00e9ristiques qu'un trieur peut \u00e9valuer est vaste, comme le montre un aper\u00e7u de fonctions d'inspection visuelle<\/a> comme on le voit clairement, m\u00eame si seul un sous-ensemble d'entre elles est pertinent pour un produit donn\u00e9. La contrainte d\u00e9terminante est d'ordre temporel plut\u00f4t que calculatoire. La d\u00e9cision doit \u00eatre prise avant que l'objet ne quitte la fen\u00eatre d'\u00e9jection, ce qui, \u00e0 une vitesse de 3 m\u00e8tres par seconde et avec un \u00e9cart de 300 millim\u00e8tres, correspond \u00e0 un d\u00e9lai total d'environ 100 millisecondes.<\/p>\n\n\n\n

\u00c9tape 4 : \u00c9jection et v\u00e9rification<\/h3>\n\n\n\n

Les produits rejet\u00e9s sont d\u00e9vi\u00e9s vers une goulotte distincte tandis que les produits accept\u00e9s poursuivent leur trajectoire. Deux indicateurs d\u00e9crivent le r\u00e9sultat et s'opposent l'un \u00e0 l'autre. L'efficacit\u00e9 d'\u00e9limination correspond \u00e0 la proportion de v\u00e9ritables d\u00e9fauts \u00e9ject\u00e9s. Le \u00ab carryover \u00bb, ou perte de produits conformes, correspond \u00e0 la proportion de produits acceptables \u00e9limin\u00e9s par erreur. Le resserrement des seuils am\u00e9liore le premier indicateur et aggrave le second ; ainsi, un trieur n\u2019est correctement r\u00e9gl\u00e9 que lorsque ces deux indicateurs se situent dans les limites de la tol\u00e9rance commerciale d\u00e9finie pour ce produit.<\/p>\n\n\n\n

\"how<\/figure>\n\n\n\n

Le syst\u00e8me de cam\u00e9ras : ce que la machine peut voir<\/h2>\n\n\n\n

La cam\u00e9ra d\u00e9termine ce que la machine est capable de d\u00e9tecter, et ses deux caract\u00e9ristiques principales sont la r\u00e9solution spatiale, qui d\u00e9finit le plus petit d\u00e9faut visible, et la fr\u00e9quence de balayage, qui d\u00e9termine la vitesse de d\u00e9filement avant que le flou de mouvement ne vienne alt\u00e9rer cette r\u00e9solution.<\/strong><\/p>\n\n\n\n

Cam\u00e9ras monochromes, couleur et trichromatiques<\/h3>\n\n\n\n

Une cam\u00e9ra monochrome enregistre uniquement la luminosit\u00e9. Elle est peu co\u00fbteuse et rapide, et s\u2019av\u00e8re efficace lorsque les d\u00e9fauts pr\u00e9sentent de fortes diff\u00e9rences de luminance, comme des grains sombres dans du riz blanc ou des \u00e9clats clairs sur du plastique fonc\u00e9. Une cam\u00e9ra couleur d\u00e9compose la lumi\u00e8re en canaux rouge, vert et bleu et effectue un tri par teinte, ce qui est important pour d\u00e9tecter les grains br\u00fbl\u00e9s et les r\u00e9sidus de broyage de couleurs m\u00e9lang\u00e9es. Les mod\u00e8les trichromatiques \u00e9tendent ce principe \u00e0 des bandes s\u00e9lectionn\u00e9es pour une application sp\u00e9cifique, incluant parfois le proche infrarouge ou l\u2019ultraviolet, de sorte que l\u2019ensemble des canaux optimise la s\u00e9paration plut\u00f4t que de reproduire la vision humaine.<\/p>\n\n\n\n

R\u00e9solution, d\u00e9bit de ligne et plus petit d\u00e9faut d\u00e9tectable<\/h3>\n\n\n\n

La r\u00e9solution s'exprime de pr\u00e9f\u00e9rence en millim\u00e8tres par pixel plut\u00f4t qu'en m\u00e9gapixels, car le champ de vision est tout aussi important que la taille du capteur. Un d\u00e9faut doit mesurer entre deux et trois pixels de large pour pouvoir \u00eatre segment\u00e9 de mani\u00e8re coh\u00e9rente.<\/p>\n\n\n\n

Largeur du capteur<\/td>Champ de vision<\/td>Taille des pixels<\/td>D\u00e9faut minimal fiable<\/td><\/tr>
2048 px<\/td>400 mm<\/td>0.20 mm<\/td>0.4 to 0.6 mm<\/td><\/tr>
4096 px<\/td>600 mm<\/td>0.15 mm<\/td>0.3 to 0.45 mm<\/td><\/tr>
8192 px<\/td>800 mm<\/td>0.10 mm<\/td>0.2 to 0.3 mm<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

A faster line needs a shorter exposure to avoid blur, and a shorter exposure needs more light, so lighting and camera choice become one problem. Mounting several cameras on the same line, as in contr\u00f4le visuel des six faces<\/a>, is the usual answer for small parts whose critical features sit on different faces.<\/p>\n\n\n\n

How a Colour Sorter Differs from a Spectral Sorter<\/h3>\n\n\n\n

A colour sorter images visible reflectance and separates on hue, saturation and shape. It performs well when the defect looks different and fails when it does not, as with black plastic in black plastic. Spectral sensing closes that gap by interrogating composition. Near-infrared instruments measure absorption at molecular vibration bands, so polymers return distinct signatures regardless of colour, a basis described in references on near-infrared spectroscopy<\/a>.<\/p>\n\n\n\n

Sensor<\/td>Signal measured<\/td>Limitation<\/td><\/tr>
RGB camera<\/td>Visible reflectance<\/td>Cannot see composition<\/td><\/tr>
NIR, 900 to 1700 nm<\/td>Molecular absorption<\/td>Needs calibration, weak on dark surfaces<\/td><\/tr>
Laser<\/td>Reflectance at one wavelength<\/td>Single spectral point<\/td><\/tr>
X-ray transmission<\/td>Density attenuation<\/td>Cost, radiation handling<\/td><\/tr>
Hyperspectral<\/td>Hundreds of narrow bands<\/td>Data volume, slower throughput<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\"optical<\/figure>\n\n\n\n

Lighting: The Subsystem That Decides Contrast<\/h2>\n\n\n\n

Lighting determines whether a defect produces a usable signal at all, and it is usually the cheapest performance lever, since a change of geometry or wavelength can outperform a camera upgrade.<\/strong><\/p>\n\n\n\n

Imaging measures light, so the geometry that delivers light to the part controls what the sensor can distinguish. The configurations used across machine vision<\/a> translate directly into sorting.<\/p>\n\n\n\n

Geometry<\/td>Light delivery<\/td>What it reveals<\/td><\/tr>
Bright field<\/td>Direct, near the camera axis<\/td>Surface colour, print, general appearance<\/td><\/tr>
Dark field<\/td>Shallow grazing angle<\/td>Scratches, cracks, burrs, raised contamination<\/td><\/tr>
Backlight<\/td>Diffuse panel behind the part<\/td>Silhouette, holes, missing features<\/td><\/tr>
Diffuse dome<\/td>Soft, omnidirectional<\/td>Shiny or curved parts prone to glare<\/td><\/tr>
Coaxial<\/td>Through the lens axis<\/td>Flat reflective surfaces such as polished metal<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Many defects are geometric rather than chromatic. A hairline crack on a machined fastener may be invisible under frontal illumination yet appear as a bright line under grazing light, because the crack scatters light toward the lens while the surrounding surface reflects it away.<\/p>\n\n\n\n

Contrast also depends on wavelength. Blue light on a yellow flaw suppresses the background and darkens the flaw, while a bandpass filter that rejects ambient light can lift a marginal contrast ratio into a robust one. Ultraviolet excitation adds a third option, since some adhesives, oils and biological contaminants fluoresce under UV while the substrate does not.<\/p>\n\n\n\n

LED arrays have displaced halogen sources because they switch in microseconds, letting a strobe pulse synchronise with the camera shutter and freeze motion. Two problems remain. Heat degrades output over thousands of hours, so intensity must be referenced against a white target at intervals, and dust or condensation on the window attenuates light, producing a slow loss of accuracy that operators notice only when rejection rates drift.<\/p>\n\n\n\n

Ejection: Turning a Decision into Physical Separation<\/h2>\n\n\n\n

The ejection system converts a classification into a physical outcome, and its performance is bounded by valve response time, pulse duration and the accuracy of the delay calculated between detection and arrival at the nozzle.<\/strong><\/p>\n\n\n\n

Detection accuracy is worthless without timing accuracy. When the camera identifies a defect, the object keeps travelling toward the ejector array, so the controller must compute the transit delay from measured velocity and distance and then open the correct valve for the correct interval.<\/p>\n\n\n\n

Param\u00e8tres<\/td>Typical range<\/td>Effect of getting it wrong<\/td><\/tr>
Temps de r\u00e9ponse de la vanne<\/td>0.5 to 1 ms<\/td>Late pulse misses the target<\/td><\/tr>
Pulse duration<\/td>5 to 20 ms<\/td>Too short, incomplete deflection; too long, extra carryover<\/td><\/tr>
Detection to ejection distance<\/td>200 to 400 mm<\/td>Sets timing budget and speed sensitivity<\/td><\/tr>
Pression atmosph\u00e9rique<\/td>3 to 7 bar<\/td>Low pressure under-deflects; excess disturbs neighbours<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

At 3 metres per second an object moves 3 millimetres every millisecond, so a valve opening one millisecond late strikes the trailing edge of the target or clips its neighbour. Nozzle bore and angle then decide how much of the pulse lands where it should. Small bores conserve air but lose force over distance, while large bores deflect reliably but disturb nearby product.<\/p>\n\n\n\n

Air quality is the most neglected variable in the machine. Moisture, oil carryover and particulates clog orifices, cause valves to under-fire and contaminate product, so a dryer, coalescing filter and dew point checks are not optional on a three-shift line.<\/p>\n\n\n\n

Pneumatic ejection dominates because it is fast and non-contact, but heavy items need air volumes that make consumption expensive, and fragile items can be damaged by the pulse. Flap gates, swing arms and retracting belts are used where objects are large, high value, or need to be placed rather than deflected, at the cost of cycle time.<\/p>\n\n\n\n

\"UGO-FB-I<\/figure>\n\n\n\n

Types of Optical Sorting Machines<\/h2>\n\n\n\n

Sorters are classified along two independent axes, the transport platform that presents material to the optics and the sensor set that performs the inspection, and the right combination is dictated by product geometry rather than budget.<\/strong><\/p>\n\n\n\n

Plateforme<\/td>Product presentation<\/td>Best suited to<\/td><\/tr>
Belt<\/td>Monolayer on a flat or inclined belt<\/td>Irregular or fragile items needing stable orientation<\/td><\/tr>
Chute<\/td>Free fall through a vertical channel<\/td>Free-flowing granular material at high capacity<\/td><\/tr>
Plateau tournant en verre<\/td>Spread on a rotating transparent disc<\/td>Small parts inspected from above and below<\/td><\/tr>
Plaque d'indexation<\/td>Located in pockets at fixed pitch<\/td>Precision parts needing repeatable position<\/td><\/tr>
Ceinture en acier<\/td>Metal carrier with high wear resistance<\/td>Abrasive, hot or heavy material<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Single-sensor machines are simpler to calibrate and faster to decide. Multi-sensor machines fuse colour with near-infrared and laser data, which sharpens discrimination but adds complexity, because channels must be spatially registered and their thresholds interact.<\/p>\n\n\n\n

Where Optical Sorting Machines Are Used<\/h2>\n\n\n\n

Optical sorting is applied wherever a stream contains items that must be separated by appearance or composition, from bulk agricultural commodities to micrometre-scale electronic components.<\/strong><\/p>\n\n\n\n

In food and agriculture, sorters remove discoloured kernels, stones and plastic, and grade by colour and size. In recycling they separate polymers by resin type and glass by colour. In mining they pre-concentrate ore before grinding, saving energy downstream.<\/p>\n\n\n\n

Precision manufacturing is the fastest-growing segment and imposes requirements bulk sorters rarely face. Fasteners are checked for thread damage and plating faults, rubber seals for flash and dimensional drift, and electronic components across multiple faces at high count rates, which is why electronics industry sorting<\/a> relies on turntable and index-plate platforms with multi-view optics.<\/p>\n\n\n\n

\"Solution<\/figure>\n\n\n\n

FAQ<\/h2>\n\n\n\n

Can an optical sorting machine separate materials that look identical?<\/strong> Yes, provided a sensor sees a property other than colour. Near-infrared channels distinguish polymers by molecular absorption, laser sensors detect transparent contaminants that share the product colour, and X-ray transmission separates by density. The limit is contrast rather than principle: if two materials return indistinguishable signals in every band, no software will separate them reliably.<\/p>\n\n\n\n

Can one machine apply several sorting criteria at once?<\/strong> Most sorters evaluate multiple features in a single pass and combine colour, shape, texture and spectral rules into one decision. Throughput cannot be combined freely, however. Each added criterion narrows the accept window, which usually raises carryover, and sensor fusion adds calibration overhead.<\/p>\n\n\n\n

How much do temperature, humidity and vibration affect stability?<\/strong> More than most specifications admit. LED output falls as junction temperature rises, so an uncorrected warm-up period shifts apparent colour. Condensation on cooled windows scatters light where cold product meets humid air. Vibration and speed fluctuation blur images and shift ejection timing. Enclosure cooling, air purging and encoder-synchronised triggering address all three.<\/p>","protected":false},"excerpt":{"rendered":"

Inspection is now the tightest constraint on many high-volume lines. A plant running several tonnes per hour cannot station enough […]<\/p>\n","protected":false},"author":16,"featured_media":6316,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[18,17],"tags":[],"class_list":["post-7717","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-knowledge-news","category-product-news"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/posts\/7717","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/users\/16"}],"replies":[{"embeddable":true,"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/comments?post=7717"}],"version-history":[{"count":1,"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/posts\/7717\/revisions"}],"predecessor-version":[{"id":7722,"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/posts\/7717\/revisions\/7722"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/media\/6316"}],"wp:attachment":[{"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/media?parent=7717"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/categories?post=7717"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.unitecho.net\/fr\/wp-json\/wp\/v2\/tags?post=7717"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}