Common GCP Mistakes in Drone Survey Malaysia 15 Ground Control Point Errors to Avoid

Common GCP Mistakes in Drone Survey Malaysia 鈥 15 Ground Control Point Errors to Avoid Content Common Ground Control Point (GCP) Mistakes in Drone Survey Malaysia 15 GCP Errors That Can Affect UAV Mapping, Photogrammetry, Construction, Earthwork & Topographic Survey Ground Control Points can significantly strengthen a drone-mapping workflow. But GCPs can also create problems when they are: Poorly positioned Measured incorrectly Too difficult to see Imported using the wrong coordinate system or Used incorrectly during processing One important principle comes before everything else: USING GCPs DOES NOT AUTOMATICALLY MAKE A DRONE SURVEY ACCURATE. A GCP is only useful when the physical target, surveyed coordinate, image identification and processing role all correspond correctly. QUICK ANSWER: The most serious GCP mistakes are often not about the target itself. They are errors involving coordinates, reference systems, placement, movement, image visibility, GCP/Check Point confusion and processing. Mistake 1 鈥 Thinking the GCP Target Creates Accuracy You buy a professional-looking black-and-white target. You place it on the ground. You fly the drone. Does that automatically improve the survey? NO. The target itself is primarily a visible reference marker. It still needs: A reliable surveyed coordinate Correct reference point Appropriate distribution Good image visibility Correct processing TARGET 鈮 ACCURACY COMPLETE WORKFLOW = SURVEY CONTROL Mistake 2 鈥 Surveying the Wrong Part of the Target Suppose your target has a clear centre. The GNSS surveyor measures: THE CORNER But the photogrammetry operator marks: THE CENTRE Now the ground coordinate and image point represent two different physical locations. FIELD POINT MUST EQUAL IMAGE POINT. Before starting a project, define exactly which part of the target represents the GCP coordinate. Mistake 3 鈥 Moving the Target After Measuring It This is one of the easiest mistakes to understand鈥攁nd still one of the easiest to make. Workflow: Place target 鈫 Survey target 鈫 Someone moves target 鈫 Drone captures new location The coordinate still represents the old location. ONCE SURVEYED, DO NOT MOVE THE TARGET UNTIL THE REQUIRED IMAGERY HAS BEEN CAPTURED. Possible causes include: Workers Vehicles Wind Excavators Animals Poor securing Mistake 4 鈥 Choosing a GCP That Is Too Small to Identify Reliably A 50脳50cm target is highly portable. But portability is not useful if the target becomes difficult to identify at the planned GSD. Target selection should consider: GSD Camera Flight parameters Ground contrast Image quality Site conditions THE TARGET MUST BE LARGE ENOUGH IN THE IMAGE FOR THE INTENDED REFERENCE POINT TO BE IDENTIFIED RELIABLY. Mistake 5 鈥 Assuming Bigger GCP = Better Accuracy The opposite mistake also occurs. A customer thinks: 鈥150cm must be three times better than 50cm.鈥 Not so. MTM Precision's three practical size categories are: 50脳50cm 鈥 Compact 100脳100cm 鈥 General Purpose 150脳150cm 鈥 Large Visual Target The main difference is physical visual footprint. 150cm TARGET 鈮 AUTOMATICALLY HIGHER SURVEY ACCURACY If 100cm is already clearly identifiable, increasing target size does not automatically improve every other component of the workflow. Mistake 6 鈥 Choosing GCP Size Only by Flight Altitude Users often ask: 鈥淚 fly at 100m. What size target do I need?鈥 Altitude is only part of the story. Different drones at the same altitude can have different: Sensors Image resolutions Lenses Focal lengths Therefore: SAME ALTITUDE 鈮 NECESSARILY SAME GSD Use expected image scale/GSD and actual visibility conditions when selecting target size. Mistake 7 鈥 Putting All GCPs in One Area Imagine a large construction project. You have eight excellent GCPs. But all eight are located beside the site office. MORE POINTS DO NOT AUTOMATICALLY COMPENSATE FOR POOR DISTRIBUTION. Control should appropriately represent the geometry of the mapped area. DISTRIBUTION MATTERS 鈥 NOT JUST COUNT. Mistake 8 鈥 Blindly Following 鈥5 GCP鈥 Search online and you may repeatedly see: 鈥淯se 5 GCPs.鈥 Five may be suitable in some project designs. But: 5 IS NOT A UNIVERSAL NUMBER FOR EVERY DRONE SURVEY. A: Small rectangular site 10km highway Deep quarry Large plantation High-rise construction site do not have identical geometry. Mistake 9 鈥 Using 鈥淥ne GCP Per Hectare鈥 as a Universal Rule Another shortcut is: 鈥淥ne GCP per hectare.鈥 Again: THERE IS NO UNIVERSAL PER-HECTARE RULE THAT FITS EVERY PROJECT. GCP requirements depend on factors such as: Geometry Terrain Flight design Required accuracy RTK/PPK workflow Survey specification Verification strategy Mistake 10 鈥 Ignoring Elevation This is especially dangerous for: Quarries Earthwork Roads Hilly terrain Land development A project is not merely: X + Y It is: X + Y + Z If all control is concentrated at similar elevation while the project has major height variation, the control design may not appropriately represent the site's 3D geometry. Mistake 11 鈥 Wrong Coordinate Reference System This can create a much bigger problem than buying the wrong target size. You can have: Perfect GCP targets Excellent GNSS measurements Sharp drone images and still produce incorrect project coordinates. Possible issues include: Wrong datum Wrong projection Wrong project grid Wrong units Wrong transformation Wrong vertical reference GOOD MEASUREMENT + WRONG CRS = WRONG PROJECT RESULT Mistake 12 鈥 X and Y Reversed During Import Imagine your coordinate file contains: Easting, Northing, Elevation But the software import is configured for: Northing, Easting, Elevation The consequences can be dramatic. DO NOT ASSUME THE SOFTWARE KNOWS YOUR COLUMN ORDER. Before processing: CHECK X / Y / Z FORMAT. Mistake 13 鈥 Ignoring Vertical Reference A particularly important problem for engineering applications. The project may require a particular elevation reference. But the imported GNSS data may use another height representation. The horizontal map may appear convincing. The vertical result may not match the intended project reference. CHECK THE Z REFERENCE BEFORE CALCULATING EARTHWORK OR VOLUME. Mistake 14 鈥 Treating RTK Fixed as Proof of Correct Coordinates GNSS rover says: FIXED Good. But this does not prove that everything else is correct. Potential issues still include: Wrong base coordinate Wrong CRS Incorrect pole height Poor point identification Wrong correction setup Data-entry error RTK FIXED 鈮 GUARANTEED CORRECT PROJECT COORDINATE Mistake 15 鈥 Using Every Surveyed Target as a GCP Suppose you survey ten targets. Many users automatically import all ten as GCPs. But then: What will independently verify the resulting mapping? Depending on project methodology, reserve appropriate targets as Check Points. GCP = CONTROL CHECK POINT = VERIFY Bonus Mistake 鈥 Calling GCP Residuals 鈥淎ccuracy鈥 Your software reports very small GCP residuals. Does that prove the entire project independently meets the required accuracy? NOT BY ITSELF. The GCPs participated in controlling the model. Independent Check Points provide a different evaluation. GCP RESIDUAL 鈮 INDEPENDENT CHECKPOINT ERROR Bonus Mistake 鈥 Hiding the Check Point Result This is a poor quality-control habit. Suppose: GCP residuals look excellent but independent Check Points show larger discrepancies. Do not simply ignore the Check Points because the GCP report looks better. CHECK POINTS EXIST TO TEST THE RESULT. Investigate why the discrepancy exists. Bonus Mistake 鈥 Putting Check Points Beside the GCPs If all verification points are concentrated around the same small area as the control: They may tell you relatively little about performance elsewhere. Check Point distribution should also be designed thoughtfully. Bonus Mistake 鈥 Target Hidden by a Vehicle This happens easily on construction sites. You survey the GCP. Twenty minutes later: A lorry parks over it. The drone flies. The target is gone from the imagery. Before take-off: CHECK EVERY REQUIRED GCP AGAIN. Bonus Mistake 鈥 GCP Covered by Quarry Dust Quarry targets may be visible when first deployed. Later: Crusher dust Truck movement Wind can reduce visibility. SURVEYED DOES NOT MEAN STILL VISIBLE. Check before flight. Bonus Mistake 鈥 GCP Hidden by Plantation Vegetation A perfectly measured target under an oil palm canopy is still a poor image target if the camera cannot see it. THE DRONE MUST SEE THE GCP. Do not confuse GNSS accessibility with aerial visibility. Bonus Mistake 鈥 Putting GCP on Moving Ground Avoid establishing important portable GCPs on: Active stockpiles Loose movable material Temporary platforms Vehicles Movable boards THE PHYSICAL REFERENCE POSITION MUST REMAIN STABLE. Bonus Mistake 鈥 Unsafe Placement A theoretically ideal GCP position is not worth an accident. Never place targets where field staff are exposed unnecessarily to: Traffic Heavy machinery High walls Unstable slopes Active excavation Unsafe roadways SAFETY > PERFECT GCP GEOMETRY Design control around safe access. Bonus Mistake 鈥 Assuming GCPs Fix Poor Drone Images Suppose your images have: Motion blur Poor overlap Missing coverage Strong glare Poor focus Can adding more GCPs repair everything? NO. GCPs provide control. They cannot recreate image information that was never captured properly. Bonus Mistake 鈥 Assuming GCPs Can See Through Trees This is particularly relevant for plantation and forest work. Ordinary RGB photogrammetry captures visible surfaces. If dense vegetation blocks the ground: MORE GCPs DO NOT MAKE THE CAMERA SEE THROUGH THE CANOPY. Different sensors or survey methods may be required depending on the deliverable. Bonus Mistake 鈥 Assuming a Good-Looking Orthomosaic Is Accurate A drone map can look: Sharp Colourful Detailed Professional and still have coordinate problems. VISUAL QUALITY 鈮 POSITIONAL ACCURACY Use appropriate ground verification. Bonus Mistake 鈥 Assuming Smooth Contours Mean Accurate Terrain Photogrammetry software can generate very smooth, attractive contour lines. But contour appearance does not prove the underlying terrain model is correct. VERIFY THE SURFACE MODEL BEFORE TRUSTING THE CONTOURS. Bonus Mistake 鈥 Wrong Stockpile Base For volume surveys: Excellent GCPs cannot solve an incorrectly defined base surface. GOOD CONTROL + WRONG BASE = POTENTIALLY WRONG VOLUME Stockpile measurement requires a complete methodology. Bonus Mistake 鈥 Changing Control Between Repeat Surveys Consider monthly earthwork mapping. January uses one reference. February uses another. March uses another. Now you compare surface changes. Some apparent change may come from reference inconsistency rather than actual earthwork. CONSISTENT CONTROL IS IMPORTANT FOR REPEAT MONITORING. Bonus Mistake 鈥 No Field Record A team surveys 12 targets. Back at the office: 鈥淲hich one was CP03?鈥 鈥淲as this GCP07 or GCP08?鈥 鈥淲hich centre did we measure?鈥 Avoid this. Record: Point ID Role Coordinate Reference-point definition Relevant field notes GOOD FIELD DOCUMENTATION PREVENTS OFFICE GUESSING. The Correct GCP Workflow A better workflow is: 1. PLAN THE CONTROL 鈫 2. SELECT APPROPRIATE TARGET SIZE 鈫 3. DEPLOY IN SAFE, VISIBLE LOCATIONS 鈫 4. SURVEY THE EXACT REFERENCE POINT 鈫 5. CONFIRM CRS & VERTICAL REFERENCE 鈫 6. KEEP TARGETS STABLE 鈫 7. FLY WITH APPROPRIATE IMAGE COVERAGE 鈫 8. MARK THE SAME REFERENCE POINT 鈫 9. PROCESS CORRECTLY 鈫 10. VERIFY WITH INDEPENDENT CHECK POINTS GCP QUALITY IS A PROCESS, NOT A PIECE OF BLACK-AND-WHITE MATERIAL. 50cm, 100cm or 150cm 鈥 Avoid the Wrong Buying Decision 50 脳 50cm COMPACT Choose where planned GSD provides sufficient reliable visibility. 100 脳 100cm GENERAL PURPOSE For many applications: 100脳100cm is our practical starting option. 150 脳 150cm LARGE VISUAL TARGET Choose when additional visual footprint is beneficial. Remember: BIGGER TARGET 鈮 AUTOMATICALLY BETTER ACCURACY Quick Pre-Flight GCP Checklist Before take-off: 鉁 Correct target locations 鉁 Correct target centres surveyed 鉁 Coordinates recorded 鉁 CRS confirmed 鉁 Vertical reference confirmed 鉁 GCP / Check Point roles confirmed 鉁 Targets have not moved 鉁 Targets remain visible 鉁 No vehicles blocking targets 鉁 No vegetation covering targets 鉁 Target size suitable for expected GSD 鉁 Drone flight covers all required targets FIVE MINUTES OF FIELD CHECKING CAN SAVE HOURS OF REPROCESSING OR A RETURN SITE VISIT. Frequently Asked Questions What is the biggest GCP mistake in drone surveying? There is no single mistake, but wrong coordinates or coordinate reference systems can undermine the entire control workflow even when the physical targets are excellent. Should I use as many GCPs as possible? NO. Use an appropriate number and distribution according to project requirements and reserve independent Check Points where needed. Are five GCPs enough? Sometimes they may be appropriate; sometimes not. There is no universal five-GCP rule. Does a larger GCP improve accuracy? Not automatically. A larger target primarily provides greater visual footprint. Can I move a GCP after measuring it? Not before the relevant imagery is captured if you intend to use the previously measured coordinate for that target position. Are GCP residuals enough to prove accuracy? No. Independent Check Points provide separate verification of the mapping at locations not used as adjustment control. Buy Ground Control Point Targets in Malaysia MTM Precision supplies reusable Ground Control Point Targets for professional Drone Survey and UAV Mapping workflows in Malaysia. Available Sizes 50 脳 50cm 鈥 COMPACT 100 脳 100cm 鈥 GENERAL PURPOSE 150 脳 150cm 鈥 LARGE VISUAL TARGET Suitable for: DRONE SURVEY UAV PHOTOGRAMMETRY GNSS RTK TOTAL STATION CONSTRUCTION EARTHWORK QUARRY STOCKPILE ROAD & HIGHWAY PLANTATION TOPOGRAPHIC MAPPING GCP & CHECK POINT WORK Contact MTM Precision MTM Precision Sdn Bhd Website: www.mtmpre.com.my Email: mtmpre@yahoo.com WhatsApp: 016-660 7346 Showroom & Service Centre No. 29-1 & 29-2, Jalan Bandar 18, Pusat Bandar Puchong, 47160 Puchong, Selangor, Malaysia. Supplying GCP targets and professional surveying equipment throughout Malaysia, including Selangor, Kuala Lumpur, Johor, Penang, Perak, Negeri Sembilan, Melaka, Pahang, Kedah, Perlis, Terengganu, Kelantan, Sarawak and Sabah. 19/30 completed.