BIGST_TOPIC_07_LASER_QandA_10202026.pdf

UNCLASSIFIED

Item Category Questions Government Answers

1 Technical Phase 1 signatures and classification scheme. What form will the Phase 1 3D signatures and preliminary classification scheme take: labeled ATL03 photon subsets, point clouds (*.LAZ/*.XYZ), derived feature tables, or source code?

Roughly how many objects and object classes are included, split between above-ground and subsurface? Do the ground-truth records include object position, height or least depth, and their uncertainties? Can a data dictionary or sample be released before proposals are due?

Phase 1 signatures and classification scheme.

1A) What form will the Phase 1 3D signatures and preliminary classification scheme take: labeled

ATL03 photon subsets, point clouds

(*.LAZ/*.XYZ), derived feature tables, or source code? Very simple label classification will be provided for derived feature and prototype algorithms in python.

The code would require refining.

1B) Roughly how many objects and object classes are included, split between above-ground and subsurface? For Phase 1 above-ground, we are starting with 5- 11 feature types and for subsurface, 4 feature types. However, there are over 300 types of interest for above-ground.

1C) Do the ground-truth records include object position, height or least depth, and their uncertainties? Lat/ long and height yes and baseline uncertainty for ground truth sensor types/source

1D) Can a data dictionary or sample be released before proposals are due? No

2 Technical Allowed inputs. May offerors use

NASA's ATL24 near-shore bathymetry product, along with ATL08 and ATL13, as inputs or cross-checks alongside

SCuBA? Or should subsurface processing rely only on ATL03 and

SCuBA? Will SCuBA be delivered as source code (Python or MATLAB), and can code derived from it be included in deliverables provided with Government

Purpose Rights?

2A) Allowed inputs. May offerors use NASA's

ATL24 near-shore bathymetry product, along with

ATL08 and ATL13, as inputs or cross-checks alongside SCuBA? Or should subsurface processing rely only on ATL03 and SCuBA? ATL24 is only available to 2024 and we are looking at new coordinated collects requiring us to run SCuBA and other algorithms. Depending on data latency of

ATL08/ ATL13 they may be applicable but ATL24 is not. Note most of our AOIs will be new collects from the space lidars so majority of the outputs would be generated using ATL03. Also we need to use the base resolution (time and space) and no denoising products to see the objects so ATL08 / ATL13 may not work depending on the object.

2B) Will SCuBA be delivered as source code

(Python or MATLAB), and can code derived from it be included in deliverables provided with

Government Purpose Rights? Yes, if needed but end products will be provided with derived products and object classification code.

3 Technical Reference data, AOIs, and metrics.

Which open reference datasets has the

Government already approved for comparison (for example USGS 3DEP, NOAA NGS topobathy lidar, NOAA

ENC and AWOIS/Wrecks and

Obstructions, FAA Digital Obstacle

File)? Will the Government hold back blind AOIs for acceptance testing? How should PD and PE be computed (per object or per track crossing; PE as the share of reported detections that are false), and what horizontal matching tolerance applies? Is "cm-scale" above-ground accuracy meant per object height or as an aggregate statistic? The PRS lists the PD AQL as "> 25%" while the metrics table says "> 75%". Please confirm that 75% is intended.

3A) Reference data, AOIs, and metrics. Which open reference datasets has the Government already approved for comparison (for example USGS

3DEP, NOAA NGS topobathy lidar, NOAA ENC and AWOIS/Wrecks and Obstructions, FAA Digital

Obstacle File)? Will the Government hold back blind AOIs for acceptance testing? Government will provide ground truth data, however, we welcome any additional sources that does NOT require new collection and will further enhance this effort.

3B) How should PD and PE be computed (per object or per track crossing; PE as the share of reported detections that are false), and what horizontal matching tolerance applies? Is "cm-scale" above-ground accuracy meant per object height or as an aggregate statistic? The PRS lists the

PD AQL as "> 25%" while the metrics table says

"> 75%". Please confirm that 75% is intended.

Should be PD> 75% with a goal of ±1 m elevation accuracy.

4 Technical Collections under this contract. Does

"collections performed under this contract" mean new ICESat-2 tasking

(for example, targeted off-pointing requests through NASA), or only retrieving existing ICESat-2 and GEDI archive data for designated AOIs? What turnaround should offerors plan for the

Government's authorization of discovered objects?

4. Collections under this contract. Does "collections performed under this contract" mean new ICESat-

2 tasking (for example, targeted off-pointing requests through NASA), or only retrieving existing

ICESat-2 and GEDI archive data for designated

AOIs? What turnaround should offerors plan for the Government's authorization of discovered objects? New ICESat-2 & GEDI data will be provided for most of the AOIs. For some archive data can be used for above ground objects.

5 Technical TRL 4 Starting Point

The Topic Call states that the project will begin at TRL 4 and progress to TRL 5, while also stating that prior efforts produced ground-truth data, space-lidar-derived 3D signatures, and preliminary object classifications that will be provided to the selected performer. Does the Government consider these prior results and Government-furnished technical artifacts to constitute the TRL

4 starting baseline for the LASER effort, with the selected performer responsible for developing, training, automating, and refining the task-specific AI/ML capability toward TRL 5; or must an

Offeror demonstrate at Abstract submission that its own proposed space-lidar AI/ML object-detection capability has independently achieved TRL 4?

5. TRL 4 Starting Point

The Topic Call states that the project will begin at

TRL 4 and progress to TRL 5, while also stating that prior efforts produced ground-truth data, space-lidar-derived 3D signatures, and preliminary object classifications that will be provided to the selected performer. Does the Government consider these prior results and Government-furnished technical artifacts to constitute the TRL 4 starting baseline for the LASER effort, with the selected performer responsible for developing, training, automating, and refining the task-specific AI/ML capability toward TRL 5; or must an Offeror demonstrate at Abstract submission that its own proposed space-lidar AI/ML object-detection capability has independently achieved TRL 4?

Yes the Government-furnished technical artifacts in phase 1 to constitute the TRL 4 starting baseline for the

LASER effort.

6 Technical AI/ML Training and Validation Data

In addition to the identified Government-furnished ICESat-2 data, SCuBA algorithm, object/AOI list, preliminary classifications, and space-lidar-derived

3D signatures, will the Government provide sufficient labeled or annotated data and associated ground-truth/reference data to support training and validation of the required AI/ML object-detection models? If not, should

Offerors propose their own approach for generating training labels and/or acquiring supplemental open-source or commercial reference data?

6. AI/ML Training and Validation Data

In addition to the identified Government-furnished

ICESat-2 data, SCuBA algorithm, object/AOI list, preliminary classifications, and space-lidar-derived

3D signatures, will the Government provide sufficient labeled or annotated data and associated ground-truth/reference data to support training and validation of the required AI/ML object-detection models? If not, should Offerors propose their own approach for generating training labels and/or acquiring supplemental open-source or commercial reference data? Government will provide ground truth data, however, we welcome any additional sources that will further enhance this effort.

7 Technical Ground-Truth and Additional Data

Collection

The Topic Call references development of test methodology and data collection and requires comparison against ground truth, including ±1-meter Z accuracy for underwater targets and cm-scale accuracy for above-ground targets. Is the selected performer expected to conduct new physical field/survey data collection, or will Government-furnished and Government-approved existing reference datasets provide the primary ground truth for algorithm development and validation?

7. Ground-Truth and Additional Data Collection

The Topic Call references development of test methodology and data collection and requires comparison against ground truth, including ±1-meter Z accuracy for underwater targets and cm-scale accuracy for above-ground targets. Is the selected performer expected to conduct new physical field/survey data collection, or will

Government-furnished and Government-approved existing reference datasets provide the primary ground truth for algorithm development and validation? No new collections expected, Government-furnished and Government-approved existing reference datasets will be provided.

8 Technical Probability of Detection Metric

Section 4 and other portions of the Topic

Call specify a Probability of Detection

(PD) greater than 75 percent and

Probability of Error (PE) less than 25 percent. However, the Performance

Requirements Summary table appears to list an AQL of “PD >25%” while its corresponding performance standard states that PD is to remain above 75 percent. Please confirm that the required

PD threshold is greater than 75 percent.

8. Probability of Detection Metric

Section 4 and other portions of the Topic Call specify a Probability of Detection (PD) greater than

75 percent and Probability of Error (PE) less than

25 percent. However, the Performance

Requirements Summary table appears to list an

AQL of “PD >25%” while its corresponding performance standard states that PD is to remain above 75 percent. Please confirm that the required

PD threshold is greater than 75 percent.

Should be PD> 75% with a goal of ±1 m elevation accuracy.

9 Security Foreign-National Participation

The Topic Call states that foreign nationals are not permitted to perform unclassified work without prior written approval from the Contracting Officer or

COR. May foreign-national personnel, including university faculty or graduate researchers, be proposed in the Abstract and proposal subject to subsequent

Government approval, or must such

All non-US personnel will be evaluated during review of the proposal submission. Evaluation and consideration is not required prior to proposal submission.

approval be obtained prior to

Abstract/proposal submission?

10 Contract Eligibility. Are Phase 1 performers of this effort eligible to submit abstracts and proposals under Topic 07?

Phase 1 performers can submit an abstract for this effort. NGA will conduct a review and technical evaluation based on the submitted abstract, without reference or regard to exisitng contracts. We will provide recommendation for follow-on proposals.

11 Contract Phase 1 data. Will the Phase 1 results, signatures, and object/area lists be shared with all offerors before proposal submission, or only with the awardee after award?

NGA intends to provide data obtained from Phase 1 to

Awardee. Data collected during Phase 1 will not be share prior to contract award.

12 Contract Award instrument. The Topic Call lists

"Procurement Contract or OTA." Is a firm-fixed-price procurement contract acceptable for this effort? If an OTA is preferred, does a nonprofit research institution (university) participating to a significant extent satisfy the prototype-

OT participation condition?

NGA does not recommend a preferred contract vehicle for this topic call. The offeror should provide a proposal with recommended contract type. A firm-fixed-price is an acceptable option for recommendation by the offeror. The offeror may also provide a proposal for an OTA with a University as the prime.

NGA recommends that offerors provide a detailed proposal that identifies the recommended contract type.

It will ultimately be the Contracting Officer’s sole discretion to select the award instrument type, regardless of instrument type proposed, and to negotiate all instrument terms and conditions with selectees.

13 Technical GEDI scope. Since GEDI does not penetrate water, should offerors understand subsurface (underwater) detection to rely on ICESat-2 only, with

GEDI applied to above-ground objects?

Yes, GEDI & ICESat-2 will be used for above-ground and ICESat-2 for underwater.

14 Security Topic 07 states work is up to U//FOUO and that uncleared personnel may perform at the contractor site. Please confirm that (a) no facility clearance, personnel clearance, SCIF, or DD254 applies to this topic, and that the DD254 language in Research and Technology

Protection Attachments C-5/C-6 does not apply; and (b) the post-selection security evaluation (General Solicitation Step 5) will assess only the factors called for in the Topic 07 security instructions.

NGA confirms that work performed will not exceed

Unclassified classification and the effort does not require access to a SCIF or handling of classified data.

15 Security Research security training. The Research and Technology Plan calls for research security training to be certified in the biographical sketch at the time of submission. Does this apply at the abstract stage, or only at the full proposal stage?

This applies only to the proposal stage.

16 Technical PD and PE. The requirements section and Table 3 give a probability of detection (PD) above 75% and a probability of error (PE) below 25%.

Table 5 lists the PD acceptable quality level as "PD > 25%". Which figure governs? How are PD and PE defined and scored (for example, PE as false alarms per object, per ground-track crossing, or per kilometer of track), and what minimum object size or class applies?

1A) PD and PE. The requirements section and

Table 3 give a probability of detection (PD) above

75% and a probability of error (PE) below 25%.

Table 5 lists the PD acceptable quality level as

"PD > 25%". Which figure governs?

Should be PD> 75% with a goal of ±1 m elevation accuracy.

1B) How are PD and PE defined and scored (for example, PE as false alarms per object, per ground-track crossing, or per kilometer of track), and what minimum object size or class applies?

Will be scored per object and the minimum object size will be provided during kickoff.

17 Technical Accuracy. The requirements give ±1 meter (Z) for underwater targets and centimeter scale for above-ground targets, against ground truth. Table 5 lists "Inaccuracy range +/- 10 meters" for

"Observation of elevation visibility" and refers to horizontal visibility. Do the ±1 meter and centimeter-scale figures govern the vertical error of an object's measured elevation profile? Is horizontal position error scored, and if so against what threshold?

2A) Accuracy. The requirements give ±1 meter (Z) for underwater targets and centimeter scale for above-ground targets, against ground truth. Table 5 lists "Inaccuracy range +/- 10 meters" for

"Observation of elevation visibility" and refers to horizontal visibility. Do the ±1 meter and centimeter-scale figures govern the vertical error of an object's measured elevation profile?

Yes, the ±1 meter is for the elevation (vertical).

2B) Is horizontal position error scored, and if so against what threshold?

It will be captured and threshold defined from this effort.

18 Technical Delivery dates. The Topic Call gives different dates for the same deliverables.

Final demonstration: two weeks before the end of the period of performance

(text and Table 3), no later than one week before (Table 3), and 30 days before (Table 4). Transition plan: three months before the end (Table 3) or no later than two weeks before (Table 4).

Final report: the last business day before the final two weeks (Table 3) or the final week (Table 4). Which dates govern?

We will update the annoucement. It should be Close-

Out brief NLT one-week prior and Algorithm Demo 1-month prior.

19 Technical Data and ground truth. Will the

Government furnish the ground-truth reference data used to score PD, PE and the accuracy figures? Roughly how many surface and subsurface objects, and how many ICESat-2 track crossings of them, will be available for development and for scoring? May GEDI data be taken from NASA's public archives?

4A) Data and ground truth. Will the Government furnish the ground-truth reference data used to score PD, PE and the accuracy figures? Yes, data and ground truth will be provided from the first phase

(Phase 1).

4B) Roughly how many surface and subsurface objects, and how many ICESat-2 track crossings of them, will be available for development and for scoring? In the first phase (Phase 1) we are evaluating

~6 objects for each category, but Phase 2 will provide prototype of all objects within the same AOI. Elevation threshold will be provided and Phase 2 will extract all objects in space lidar data.

4C) May GEDI data be taken from NASA's public archives?

Yes, if there is collection of the areas/ objects of interest. However, NGA will coordinate new collections for objects not collected.

20 Technical Controlled information. Which

Government-furnished items (the

SCuBA algorithm, the object and area list, the ICESat-2 data from previous collections, the ground truth) will be marked CUI or FOUO, and which are public? I ask so that the NIST SP 800-

171 and DFARS 252.204-7012 scope can be planned correctly.

Controlled information. Which Government-furnished items (the SCuBA algorithm, the object and area list, the ICESat-2 data from previous collections, the ground truth) will be marked CUI or FOUO, and which are public?

I ask so that the NIST SP 800-171 and DFARS

252.204-7012 scope can be planned correctly.

Anything developed or used under this effort must be vetted for public release if not already in public. This includes publications and presentations. Also, compiled products built from public release products must be vetted before release to the public.

21 Technical Place of performance and platforms.

Section 13.2 says work is primarily at the contractor's facility and at 7500

GEOINT Drive, Springfield, VA. How many on-site meetings should an offeror plan for (kickoff, demonstrations, final briefing), and may the rest be held virtually? Which "NGA approved platforms" (operating system, Python or

MATLAB version, container support) should the prototype demonstration target?

6A) Place of performance and platforms. Section

13.2 says work is primarily at the contractor's facility and at 7500 GEOINT Drive, Springfield, VA. How many on-site meetings should an offeror plan for (kickoff, demonstrations, final briefing), and may the rest be held virtually?

Ideally 2 meetings, Kick-off and Close-out, however, virtual meeting options are available.

6B) Which "NGA approved platforms" (operating system, Python or MATLAB version, container support) should the prototype demonstration target? Python is preferred. Also, a containerized workflow using gitlab and openshift is preferred.

22 Technical Field test and data collection. The call says earlier efforts produced the ground truth and signatures, and its revision notes say Phase 1 requirements were removed, yet the CDRLs still include a

Field Test Collection and Report (A042) and Table 5 still has a "Data

Intercomparison collection" row (more than 80% of data collected without

The field test is not mentioned in the updated call (No.

A042).

errors, more than 90% retained). Do an offeror's own field collection, A042 and that Table 5 row still apply? I plan to build and score everything from the furnished satellite data and ground truth, with the intercomparison reports (A032 and A062) and no field collection.

23 Contract Award instrument. Does NGA prefer an

Other Transaction or a procurement contract for this topic? For a prototype

Other Transaction I read BAA Section

3.3 as requiring no cost share when every significant participant is a small business, and I would be the only participant; please say if you read it differently.

The Government is contemplating issuing either a

Procurement Contract or Other Transaction

Agreement. Offeror/Performer shall submit a recommended Award Type along with the Abstract. It will ultimately be the Contracting Officer’s sole discretion to select the award instrument type, regardless of instrument type proposed, and to negotiate all instrument terms and conditions with selectees.

24 Technical Depth regime — What maximum useful water depth (or optical-depth bound) does NGA assume for “underwater objects” and seafloor products under

LASER, given ICESat-2/GEDI limits?

Depth regime — What maximum useful water depth (or optical-depth bound) does NGA assume for “underwater objects” and seafloor products under LASER, given ICESat-2/GEDI limits?

Underwater only applies to ICESat-2 and underwater limits will vary based on water clarity.

25 Technical Object classes — Which man-made classes are in scope (surface vessels, partially submerged structures, fully submerged shallow objects), and which are explicitly out of scope?

Object classes — Which man-made classes are in scope (surface vessels, partially submerged structures, fully submerged shallow objects), and which are explicitly out of scope? The focus is on larger objects for underwater (ex. shipwrecks).

26 Technical Water clarity & sea state — Must performance be reported stratified by turbidity, bottom reflectance, and sea state, or is a single aggregate score acceptable?

Water clarity & sea state — Must performance be reported stratified by turbidity, bottom reflectance, and sea state, or is a single aggregate score acceptable?

Yes, report based on environmental conditions.

27 Technical NASA product use — Is direct use of public ICESat-2 photon products (e.g.

ATL03 / operational bathymetry products) assumed, or must offerors propose independent classification and refraction correction?

NASA product use — Is direct use of public

ICESat-2 photon products (e.g. ATL03 / operational bathymetry products) assumed, or must offerors propose independent classification and refraction correction? NGA will provide derived ATL03 derived products from Phase 1 and yes ATL03 will be used and recommended for base level product to avoid data that is averaged and denoising objects of interest.

28 Technical Output form — Preferred deliverables:

photon-level classes, along-track depths/objects, gridded seafloor, vector object hypotheses—or a defined combination aligned to foundation

GEOINT formats?

Output form — Preferred deliverables: photon-level classes, along-track depths/objects, gridded seafloor, vector object hypotheses—or a defined combination aligned to foundation GEOINT formats? Photon level and along-track, as we want to hold the base resolution as much as possible.

29 Technical Shallow-only — Confirm abyssal / deep-turbid regimes beyond reliable green-lidar penetration are out of scope.

Shallow-only — Confirm abyssal / deep-turbid regimes beyond reliable green-lidar penetration are out of scope. The focus is shallow and the limitation is based on collection ability.

30 Technical Unclassified lane — Confirm abstracts/proposals and demo materials must remain unclassified and must not rely on detailed, platform-complete depictions of sensitive combat systems.

Unclassified lane — Confirm abstracts/proposals and demo materials must remain unclassified and must not rely on detailed, platform-complete depictions of sensitive combat systems. Proposals can only be Unclassified.

31 Technical The topic says separate Phase 1 efforts, not part of this BAA, produced ground truth and space-lidar 3D signatures. Is that work complete? If not, when will its outputs be complete and available to the performer?

The topic says separate Phase 1 efforts, not part of this BAA, produced ground truth and space-lidar

3D signatures. Is that work complete? If not, when will its outputs be complete and available to the performer? Phase 1 is active now and initial results will be provided to Phase 2 Awardee during kickoff.

32 Technical Please confirm that the offeror is not expected to collect field data, deploy instruments, or arrange new ICESat-

2/GEDI collections with NASA. Two passages in the 1 October topic still point to collection. The first paragraph of the

Topic Research Opportunity Description says NGA "seeks to task the offeror with development of test methodology, data collection and novel data analysis." The

Satellite Data Provided paragraph under

Project Structure/Scope says "Offerors must submit discovered objects from collections performed under this contract to the Government for authorization."

No field collection is expected for this TOPIC call.

33 Technical What target classes and approximate object sizes should offerors plan to detect and score, above ground and submerged? Roughly how many labeled objects, areas of interest, and ICESat-

2/GEDI track crossings will be provided, and how are they split between surface and subsurface?

What target classes and approximate object sizes should offerors plan to detect and score, above ground and submerged? Roughly how many labeled objects, areas of interest, and ICESat-2/GEDI track crossings will be provided, and how are they split between surface and subsurface? Phase 1 will start with 6 use cases above ground and 6 submerged, which will be provided. Phase 2 will automate detection in granules versus focusing on 6 objects.

34 Technical What form, horizontal and vertical datums, and stated accuracy will the

Phase 1 ground truth have?

What form, horizontal and vertical datums, and stated accuracy will the Phase 1 ground truth have?

Horizontal and vertical datums will be based on

WGS84 and the aim for stated accuracy is 1m or less.

35 Technical The GFI table lists the "SCuBA algorithm" without stating its form. Will

SCuBA, and the preliminary classification and object-detection scheme, be delivered as executable code, source code, or documentation? May the performer modify, extend, and redeliver the preliminary scheme as part of the prototype?

5A)The GFI table lists the "SCuBA algorithm" without stating its form. Will SCuBA, and the preliminary classification and object-detection scheme, be delivered as executable code, source code, or documentation? Code can be provided if needed.

5B) May the performer modify, extend, and redeliver the preliminary scheme as part of the prototype? Yes

36 Technical Table 5 (PRS) lists the PD acceptable quality level as "PD > 25%," while the requirements, Table 3, and CDRL A044 state PD > 75%. Is PD > 25% a minimum acceptable (threshold) level with PD > 75% as the objective, or is

PD > 75% the required level?

Table 5 (PRS) lists the PD acceptable quality level as "PD > 25%," while the requirements, Table 3, and CDRL A044 state PD > 75%. Is PD > 25% a minimum acceptable (threshold) level with PD >

75% as the objective, or is PD > 75% the required level? Should be PD> 75% with a goal of ±1 m elevation accuracy.

37 Technical How are PD and PE defined and scored?

Is the unit of detection an object, a track crossing, or an along-track segment?

Will NGA provide held-out targets or areas for final evaluation, or should offerors propose them in the QASP?

Must thresholds be met separately for surface and subsurface targets?

How are PD and PE defined and scored? Is the unit of detection an object, a track crossing, or an along-track segment? > 75 % is acceptable for vertical detection of object.

Will NGA provide held-out targets or areas for final evaluation, or should offerors propose them in the

QASP? NGA will provide objects observed in Phase 1.

Must thresholds be met separately for surface and subsurface targets? Yes.

38 Technical For the cm-scale above-ground requirement, what quantity is scored, against what reference, and at what unit of analysis? Is horizontal location scored separately?

For the cm-scale above-ground requirement, what quantity is scored, against what reference, and at what unit of analysis? Primary reference will be airborne lidar data, however, for underwater data, we will have survey data.

Is horizontal location scored separately? Yes

39 Technical Table 4 points the new "Automate

Object detection scheme" deliverable to section A032. In the posted CDRL, A032 is the AI/ML Detection Report and describes intercomparison testing. Which content requirements apply to the automated detection deliverable?

Table 4 points the new "Automate Object detection scheme" deliverable to section A032. In the posted

CDRL, A032 is the AI/ML Detection Report and describes intercomparison testing. Which content requirements apply to the automated detection deliverable? Both with emphasis on automated detection of objects in satellite data.

40 Technical The AI/ML prototype is to be demonstrated "using NGA approved platforms," and the capability transitions to NGA at the end of the period of performance. Which environment is required for the demonstration, and which will receive the transition package? What access will the performer have to each, and when? What technical constraints apply (operating system, container support, GPU availability, network restrictions, approved software)?

A containerized workflow using gitlab and openshift is preferred.

41 Technical May government-furnished data be processed, and the demonstration conducted, in a contractor cloud environment, or is use of an NGA platform mandatory for either? Will any government-furnished data be designated

CUI?

Yes, it may be used in performer environment, but not shared with the public prior to approval. Presentation/ data sharing etc. Some data is CUI//FOUO and will be provided using the proper mechanism if necessary or not at all. This will not limit the performance.

42 Technical 12. Is the $854K a ceiling or a planning budget? Is the three-month transition period included within the 15-month period of performance, or does it follow it?

$854M is the budget ceiling and a 3-month transition is included in the 15 months.

43 Field collection. The 1 October topic text states that Phase 1 field work was removed and that ground truth comes from "Separate Phase 1 efforts, which are not part of this BAA." However, CDRL A004 still refers to "physical field data collections (utilizing the contractor's surveying and remote sensing instruments for submerged and surface objects)," and CDRL A002 still lists

"(A042) Field Test Collection and

Report." Please confirm that the contractor performs no physical field collection and needs no surveying or remote sensing instruments under this topic.

The field test is not mentioned in the updated call (No.

A042).

44 Place of performance. The topic lists the

Contractor's Facility and 7500 GEOINT

Drive, Springfield, VA. May the kickoff, the recurring status meetings and the live demonstration be held virtually, or is in-person attendance required for any of them?

Place of performance. The topic lists the

Contractor's Facility and 7500 GEOINT Drive, Springfield, VA. May the kickoff, the recurring status meetings and the live demonstration be held virtually, or is in-person attendance required for any of them? Virtual meetings and Kick-offs are an option.

45 Detection threshold. The performance table gives probability of detection as ">

25%," while the text gives 75%. Which is correct?

Detection threshold. The performance table gives probability of detection as "> 25%," while the text gives 75%. Which is correct? Should be PD> 75% with a goal of ±1 m elevation accuracy.

46 Government-furnished data. In what form and under what marking will the

Phase 1 ground truth and the area-of-interest list be provided, and may the work be performed in a contractor-owned environment that meets NIST SP

800-171?

Government-furnished data. In what form and under what marking will the Phase 1 ground truth and the area-of-interest list be provided, and may the work be performed in a contractor-owned environment that meets NIST SP 800-171?

Derived products using ICESat-2 and GEDI will be provided in lidar formats .XYZ & .LAS . The work may be performed in contractor own environment but public release of presentation and products from this effort require prior approval.

47 Phase 1 performer. Is the performer of the separate Phase 1 efforts eligible to propose under this topic? Will the Phase

1 methods, results and data be furnished to all offerors before abstracts are due, or only to the awardee? Is this topic intended as an open competition among new performers?

Phase 1 performers can submit an abstract for this effort. NGA will conduct a review and technical evaluation based on the submitted abstract, without reference or regard to exisitng contracts. We will provide recommendation for follow-on proposals.

NGA will not share methods with other parties. NGA will provide final results and data from Phase 1 to awardees only in support of successful execution of phase 2 requirements.

This Topic Call is considereed open competition for all offerors.

48 Will the government please share information on SCUBA and its data as it relates to the scope of LASER?

SCuBA data is available on the open internet.

49 Will the government please provide the specifics of what the following performance outcomes mean:

"observation of elevation visibility" and

"data intercomparison collection"

In table 5, these are the titles in cloumn 1 and the definition are in column 2.

50 What format is the preliminary classification scheme/SCuBA algorithm delivered in : source code, trained model, or documentation only?

What format is the preliminary classification scheme/SCuBA algorithm delivered in : source code, trained model, or documentation only?

SCuBA derived products are currently available on the open internet.

51 How many AOIs and what data volume should Offerors plan for?

How many AOIs and what data volume should

Offerors plan for? Phase 1 will start with 6 use cases above ground and 6 submerged, which will be provided. Phase 2 will automate detection in granules versus focusing on 6 objects.

52 Which specific open-source datasets would NGA consider "approved" ground truth for this comparison?

Which specific open-source datasets would NGA consider "approved" ground truth for this comparison? Data that provides volume 3D scan of objects of interest on a cm scale ( ex. Airborne lidar data).

53 Is the ±1m underwater accuracy threshold evaluated per-object or as an aggregate RMSE, and can NGA specify a numeric cm-scale tolerance for above-ground targets?

Is the ±1m underwater accuracy threshold evaluated per-object or as an aggregate RMSE, and can NGA specify a numeric cm-scale tolerance for above-ground targets? Per object.

54 Beyond QASP review, will NGA independently spot-check PD/PE results, or is self-reporting the sole acceptance basis?

Beyond QASP review, will NGA independently spot-check PD/PE results, or is self-reporting the sole acceptance basis? Yes

55 Is GEDI data provided as GFI, or expected to be independently sourced like the rest?

Is GEDI data provided as GFI, or expected to be independently sourced like the rest? GEDI derived products will be provided as GFI but in general raw

GEDI is an independent source.

56 Are Offerors expected to build strictly on the provided classification scheme, or may a materially different technical approach be proposed if justified?

Are Offerors expected to build strictly on the provided classification scheme, or may a materially different technical approach be proposed if justified? Offers are expected to use space lidar data to automatically extract Latitude, Longitude and elevation for objects of interest.

57 In regards to the three-dimensional (3D) signatures derived from space lidar data that will be provided:

Are there signatures for both ICESat-2 and GEDI?

"In regards to the three-dimensional (3D) signatures derived from space lidar data that will be provided:Are there signatures for both ICESat-2 and GEDI?" Yes

58 In regards to the three-dimensional (3D) signatures derived from space lidar data that will be provided:

Are there signatures for all three environments: on land objects, on sea surface objects, submerged objects?

"In regards to the three-dimensional (3D) signatures derived from space lidar data that will be provided:Are there signatures for all three environments: on land objects, on sea surface objects, submerged objects?" Yes

59 In regards to the three-dimensional (3D) signatures derived from space lidar data that will be provided:

Were these data curated as part of a previously funded research project, or were these created internally at NGA?

"In regards to the three-dimensional (3D) signatures derived from space lidar data that will be provided: Were these data curated as part of a previously funded research project, or were these created internally at NGA? Work is currently underway for Phase 1.

60 With respect to the ground truth data:

Do these cover objects of interest for all three environments: on land objects, on sea surface objects, submerged objects?

"With respect to the ground truth data: Do these cover objects of interest for all three environments:

on land objects, on sea surface objects, submerged objects?" Yes

Ground-truth data and furnished items.

The topic call states that separate Phase 1 efforts “produced both ground-truth data and space lidar-derived three-dimensional (3D) signatures,” and that the Government will supply 3D signatures at the kickoff meeting, but

Table 6 (GFI/GFD/GFP) does not list ground-truth data.

(a) Will the ground-truth measurements produced by those separate efforts be furnished to the performer and, if so, of what type (e.g., airborne lidar, sonar, or field survey) and with what stated vertical accuracy? Will Table 6 be updated to list all furnished items with format and expected date?

(b) Where ground truth is not furnished, which datasets does the Government already consider “government approved open-source reference data,” and may offerors propose others for Government approval?

(c) Approximately how many objects

(underwater vs. above-ground) will the object list cover, and are the sites in U.S.

or international waters and territory?

(d) Are the object/AOI list, the 3D signatures, or any of the ground-truth or reference datasets from the separate efforts publicly accessible now, so that offerors can design and demonstrate their approach before award? If not, will any be released with the Q&A or with abstract feedback?

a) Yes, ground truth will also be provided. The formats will be .xyz and /or .las. Detailed information will be provided after award. B) As listed in the proposal, data from NOAA and USGS lidar sites are examples approved open-source data. Offers may propose additional sources not listed as long as they are 3D volume scans of objects with cm scale resolution.

C)Phase 1 will start with 6 sites above ground and 6 sites underwater with majority in the US. Phase 2 will scale to more objects in same AOI. D) Yes example ground truth data are available on lidar open internet.

Space lidar derived 3D signatures of objects will be provided to awarded contract however SCuBA 3D signatures of seafloor with similar formats are available on NOAA website.

62 PD/PE and accuracy thresholds.

(a) Table 5 lists the PD Acceptable

Quality Level as “PD > 25%,” while the same row’s standard (“remains above

75%”) and Requirement 1 (“greater than

75%”) state 75%; please confirm that

75% governs.

(b) Will the Government specify how PD and PE are computed, including the scoring unit, the PD denominator (e.g., all listed objects, or only objects with usable ICESat-2/GEDI returns), and what PE (“Probability of detection error” in Table 5) measures, or should offerors propose these definitions in the draft

QASP?

(c) Requirement 1 requires “cm-scale accuracy for above-ground targets,” while CDRL A002 (A032, A062) states a “±1 m elevation (z) agreement” goal for “both underwater and surface target collects.” Which governs above-ground targets, including GEDI-derived results?

(d) Are the PD/PE and accuracy thresholds acceptance criteria for contract compliance, or objectives whose achieved values are to be measured and reported, and will they be assessed on

Government-held data not furnished for development?

a) Should be PD> 75% with a goal of ±1 m elevation accuracy. B) Will be scored per object and per space lidar with emphasis on elevation ( z). AOIs will be provided with known objects of interest and PD/ PE will be scored based on detects. C) ±1 m elevation (z) agreement” governs both but performance of above ground elevation is expected to exceed this. D) both

63 Collections under this contract.

The topic call tasks the offeror with

“development of test methodology, data collection and novel data analysis,” states that “Offerors must submit discovered objects from collections performed under this contract to the

Government for authorization,” and

Table 5 includes a “Data

Intercomparison collection” row (“>80% data collected without errors”).

(a) Does “collection” under this contract include requesting new ICESat-2 or

GEDI acquisitions over the AOIs, or is the effort to be performed on archived and Government-furnished data only?

(b) Table 5 also contains a performance outcome titled “Observation of elevation visibility,” whose performance standard is to “determine horizontal visibility with repeatable results” and whose

Acceptable Quality Level is an

a) This will be covered in phase 1. b) The 10 meters is an error. Should be ±1 meter accuracy for both above-ground and underwater objects. Note above ground is expected to perform well within ±1 meter.

“Inaccuracy range +/- 10 meters.” What quantity does this row measure, on what data will it be sampled, and how does it relate to the ±1 meter and cm-scale accuracy requirements in Requirement

1?

64 Status of the separate efforts.

The topic call states that the separate

Phase 1 efforts “produced” ground-truth data and 3D signatures, but also that

“prior efforts will produce object classifications for large objects” and that

“A preliminary classification and object-detection scheme ... will be provided.”

(a) Are those efforts complete? If not, when will their outputs be available relative to the 15 February 2027 estimated start? (b) Is the preliminary scheme rule-based, statistical, machine-learning, analyst-driven, or a combination, and has its detection performance been measured? A summary-level answer is sufficient.

Phase 1 outputs be provided to partners after award and kick-off. It’s a combination of stats and analyst for

Phase 1 which will determine collection performance.

Phase 2 will evaluate detection performance

65 Form and rights of the furnished

SCuBA algorithm.

Table 6 (GFI/GFD/GFP) lists the

“SCuBA algorithm” with no entry under

FORMAT/SIZE.

(a) Will it be furnished as source code, an executable or container, or processed outputs only, and with what language and documentation?

(b) Which ICESat-2 data product and release does it ingest?

(c) May the performer modify it and deliver modified versions under the

Government Purpose Rights minimum described in Requirement 2 and

“Intellectual Property Considerations”?

A) Other algorithms will be provided from Phase 1.

SCuBA will only be provided if needed. SCUBA example products are available on the open internet.

B) ATL03

C) NGA allow modification of data under government purpose right clauses. However, we would consider modification in terms of technical requirements and our use of reference data (also termed ground truth data).

Acceptance of data modification will be consider post award based discussion between NGA and Awardee.

66 Permitted data sources.

The objective is "automatic detection of manmade objects from ICESat-2 and

GEDI," and the goal is for the signatures to "serve as calibration and enhancement tools for other data sources that lack elevation information."

(a) Must detections be derived from space-lidar returns alone, or may the detection workflow use ancillary data sources (e.g., imagery, existing elevation models, or charted features) for priors, localization, or confirmation?

(b) If ancillary sources are permitted, are there restrictions on their licensing or on their use in the delivered prototype?

A) Imagery or other ancillary data are permitted.

B) If ancillary sources are permitted, NGA will need to evaluate licensing restrictions on an individual case basis. Restrictions on ancillary data does not constitute failure to meet requirements as we explore innovative ways to address our mission gaps.